14 Förslag på visualisering och återkoppling av OSA-enkäten

Code

library(tidyverse)
library(ggdist)
library(ggpp)
library(foreign)
library(readxl)
library(showtext)
library(stringr)
library(patchwork)
library(glue)
library(ggridges)
library(scales)

## Loading Google fonts (https://fonts.google.com/)
font_add_google("Noto Sans", "noto")
## Flama font with regular and italic font faces
font_add(family = "flama", 
         regular = "../fonts/Flama-Font/Flama Regular.otf", 
         italic = "../fonts/Flama-Font/Flama Italic.otf",
         bold = "../fonts/Flama-Font/FlamaBlack Regular.otf")
## Automatically use showtext to render text
showtext_auto()

### some commands exist in multiple packages, here we define preferred ones that are frequently used
select <- dplyr::select
count <- dplyr::count
recode <- car::recode
rename <- dplyr::rename

# file paths will need to have "../" added at the beginning to be able to render document

# get itemlabels
itemlabels <- read_excel("../data/Itemlabels.xlsx") %>%
  filter(str_detect(itemnr, pattern = "abk|å")) %>%
  select(!Dimension)

spssDatafil <- "2023-04-05 kl15_27 Prevent OSA-enkat.sav"

# read SurveyMonkey data
df <- read.spss(spssDatafil, to.data.frame = TRUE) %>%
  select(starts_with("q0006"), starts_with("q0009"), q0001, q0002, q0003, q0004) %>%
  rename(
    Kön = q0002,
    Ålder = q0001,
    Bransch = q0003,
    Hemarbete = q0004
  ) %>%
  na.omit()

dif.kön <- df$Kön
dif.ålder <- df$Ålder
dif.bransch <- df$Bransch
dif.hemarbete <- df$Hemarbete

df <- df %>%
  select(starts_with("q0006"), starts_with("q0009"))

names(df) <- itemlabels$itemnr

# read data for one index. Includes item level data and scored data
df.scored <- read.csv("../visualisering/arbkrv_recSCORED.csv") %>%
  select(!X)

# get labels/descriptions for all items in the current domain/index
itemlabels <- itemlabels %>%
  filter(itemnr %in% c("abk4", "abk5", "abk6", "å2", "å4", "å5"))

# add index scores to raw data, for later use when we need the actual response categories
df$score <- df.scored$score

# vector of response categories
svarskategorier <- c("Aldrig","Sällan","Ibland","Ganska ofta","Mycket ofta","Alltid")

14.1 Prevent tema

Kulörer, typsnitt osv - definierar återkommande layout-inställningar för figurer.

Code

prevent_green <- "#008332"
prevent_light_green <- "#76A100"
prevent_dark_blue <- "#003E6E"
prevent_blue <- "#005F89"
prevent_light_blue <- "#4398BA"
prevent_yellow <- "#FBB900"
prevent_red <- "#BE5014"
prevent_gray_red <- "#6C5861"
prevent_light_gray <- "#F0F0F0"
prevent_gray <- "#d3d3d3"
prevent_dark_gray <- "#3B3B3B"
prevent_turquoise <- "#009a9d"
prevent_green_comp <- "#D9ECE0"
prevent_light_green_comp <- "#DCE7BF"
prevent_dark_blue_comp <- "#BFCEDA"
prevent_blue_comp <- "#BFD7E1"
prevent_light_blue_comp <- "#D0E5EE"
prevent_yellow_comp <- "#FEEDBF"
prevent_red_comp <- "#EFD3C4"
prevent_green_contrast <- "#006632"
prevent_blue_contrast <- "#003E6E"
prevent_yellow_contrast <- "#FBD128"
prevent_red_contrast <- "#B01200"
prevent_gray_red_contrast <- "#68534E"

# manual palette creation, 7 colors
PREVENTpalette1 <- c("#6C5861", "#005F89", "#4398BA", "#76A100", "#008332", "#FBB900", "#FBD128")

# create palette with 12 colors based on Prevent colors above
PREVENTpalette2 <- colorRampPalette(colors = c("#6C5861", "#005F89", "#4398BA", "#76A100", "#008332", "#FBB900", "#FBD128"))(12)
# scales::show_col(PREVENTpalette2)

Code

theme_prevent <- function(fontfamily = "flama", axisTitleSize = 13, titlesize = 15,
                          margins = 12, axisface = "plain", stripsize = 12,
                          panelDist = 0.6, legendSize = 9, legendTsize = 10,
                          axisTextSize = 10, ...) {
  theme(
    text = element_text(family = fontfamily),
    axis.title.x = element_text(
      margin = margin(t = margins),
      size = axisTitleSize
    ),
    axis.title.y = element_text(
      margin = margin(r = margins),
      size = axisTitleSize
    ),
    plot.title = element_text(
      #family = "flama",
      face = "bold",
      size = titlesize
    ),
    axis.title = element_text(
      face = axisface
    ),
    axis.text = element_text(size = axisTextSize),
    plot.caption = element_text(
      face = "italic"
    ),
    legend.text = element_text(family = fontfamily, size = legendSize),
    legend.title = element_text(family = fontfamily, size = legendTsize),
    strip.text = element_text(size = stripsize),
    panel.spacing = unit(panelDist, "cm", data = NULL),
    legend.background = element_rect(color = "lightgrey"),
    ...
  )
}

# these rows are for specific geoms, such as geom_text() and geom_text_repel(), to match font family. Add as needed
#    update_geom_defaults("text", list(family = fontfamily)) +
#    update_geom_defaults("text_repel", list(family = fontfamily)) +
#    update_geom_defaults("textpath", list(family = fontfamily)) +
#    update_geom_defaults("texthline", list(family = fontfamily))

14.2 Gruppnivå

Nedanstående kod skapar underlag för figurerna, med sampelstorlek 7, 20 och 40.

Code

# create vector with possible values, which needs to be connected to the transformation table
# ordinal -> interval
# "score" is interval level
possible <- df.scored %>% # maybe rewrite this to just make the dataframe below
  distinct(score) %>%
  arrange(score) %>%
  pull(score)

# make a dataframe with possible values and a count variable (n) with missing data
df.possible <- possible %>%
  as.data.frame(nm = "score") %>%
  add_column(n = NA)

sampleSmall <- 7
sampleMed <- 20
sampleLarge <- 40
set.seed(1337)

Code

# pick random sample to use for visualization example
df.test7 <- df.scored %>%
  slice_sample(n = sampleSmall) %>%
  add_column(group = "Mättillfälle 1")

# get another sample for examples comparing two measurements
df.test7b <- df.scored %>%
  slice_sample(n = sampleSmall) %>%
  add_column(group = "Mättillfälle 2")

# combine data
df.compare7 <- rbind(df.test7, df.test7b)

###
# pick random sample to use for visualization example
df.test20 <- df.scored %>%
  slice_sample(n = sampleMed) %>%
  add_column(group = "Mättillfälle 1")

# get another sample for examples comparing two measurements
df.test20b <- df.scored %>%
  slice_sample(n = sampleMed) %>%
  add_column(group = "Mättillfälle 2")

# combine data
df.compare20 <- rbind(df.test20, df.test20b)

###
# pick random sample to use for visualization example
df.test40 <- df.scored %>%
  slice_sample(n = sampleLarge) %>%
  add_column(group = "Mättillfälle 1")

# get another sample for examples comparing two measurements
df.test40b <- df.scored %>%
  slice_sample(n = sampleLarge) %>%
  add_column(group = "Mättillfälle 2")

# combine data
df.compare40 <- rbind(df.test40, df.test40b)

Code

# create a dataframe with frequency counts of estimated scores, to use for size/color
df.counts7 <- df.test7 %>%
  count(score) %>%
  right_join(df.possible, # fill in with all scores possible
    by = "score"
  ) %>%
  arrange(score) %>%
  select(!n.y) %>%
  rename(n = n.x) %>% 
  mutate(N = case_when( # change missing to 1 and multiply other frequencies by 2
    is.na(n) ~ 1,
    TRUE ~ n * 2
  )) %>%
  mutate(nFactor = factor(n)) # make a categorical variable with counts

# calculate 90% confidence interval for the test sample
smallSample90ci <- sd(df.test7$score) / (sqrt(length(df.test7$score))) * 1.65
smallSample90ci2 <- sd(df.test7b$score) / (sqrt(length(df.test7b$score))) * 1.65

#####
# create a dataframe with frequency counts of estimated scores, to use for size/color
df.counts20 <- df.test20 %>%
  count(score) %>%
  right_join(df.possible, # fill in with all scores possible
    by = "score"
  ) %>%
  arrange(score) %>%
  select(!n.y) %>%
  rename(n = n.x) %>% 
  mutate(N = case_when( # change missing to 1 and multiply other frequencies by 2
    is.na(n) ~ 1,
    TRUE ~ n * 2
  )) %>%
  mutate(nFactor = factor(n)) # make a categorical variable with counts

# calculate 90% confidence interval for the test sample
medSample90ci <- sd(df.test20$score) / (sqrt(length(df.test20$score))) * 1.65
medSample90ci2 <- sd(df.test20b$score) / (sqrt(length(df.test20b$score))) * 1.65

#####
# create a dataframe with frequency counts of estimated scores, to use for size/color
df.counts40 <- df.test40 %>%
  count(score) %>%
  right_join(df.possible, # fill in with all scores possible
    by = "score"
  ) %>%
  arrange(score) %>%
  select(!n.y) %>%
  rename(n = n.x) %>% 
  mutate(N = case_when( # change missing to 1 and multiply other frequencies by 2
    is.na(n) ~ 1,
    TRUE ~ n * 2
  )) %>%
  mutate(nFactor = factor(n)) # make a categorical variable with counts

# calculate 90% confidence interval for the test sample
largeSample90ci <- sd(df.test40$score) / (sqrt(length(df.test40$score))) * 1.65
largeSample90ci2 <- sd(df.test40b$score) / (sqrt(length(df.test40b$score))) * 1.65

Code

# for comparison of 2 measurements
df.countsComp7 <- df.compare7 %>%
  group_by(group) %>%
  count(score) %>%
  right_join(df.possible, # fill in with all scores possible
    by = "score"
  ) %>%
  arrange(score) %>%
  select(!n.y) %>%
  rename(n = n.x) %>% 
  mutate(N = case_when( # change missing to 1 and multiply other frequencies by 2
    is.na(n) ~ 1,
    TRUE ~ n * 2
  )) %>%
  mutate(nFactor = factor(n)) # make a categorical variable with counts

# for comparison of 2 measurements
df.countsComp20 <- df.compare20 %>%
  group_by(group) %>%
  count(score) %>%
  right_join(df.possible, # fill in with all scores possible
    by = "score"
  ) %>%
  arrange(score) %>%
  select(!n.y) %>%
  rename(n = n.x) %>% 
  mutate(N = case_when( # change missing to 1 and multiply other frequencies by 2
    is.na(n) ~ 1,
    TRUE ~ n * 2
  )) %>%
  mutate(nFactor = factor(n)) # make a categorical variable with counts

# for comparison of 2 measurements
df.countsComp40 <- df.compare40 %>%
  group_by(group) %>%
  count(score) %>%
  right_join(df.possible, # fill in with all scores possible
    by = "score"
  ) %>%
  arrange(score) %>%
  select(!n.y) %>%
  rename(n = n.x) %>% 
  mutate(N = case_when( # change missing to 1 and multiply other frequencies by 2
    is.na(n) ~ 1,
    TRUE ~ n * 2
  )) %>%
  mutate(nFactor = factor(n)) # make a categorical variable with counts

14.3 Visualisering av ett område/index

Vi kommer antagligen att vilja ta fram en vy som ger en överblick, där resultat på samtliga områden visas under varandra. Som ett första steg tar vi fram förslag på hur ett enskild index kan visas, och i nästa steg hur vi visualiserar svaren på de frågor/items som ingår i ett index, eller på frågor som inte kan bilda ett index.

14.3.1 Förslag 1

Detta ser kanske något suboptimalt ut när det bara är ett index (eller en grupp/mätning) som visualiseras.

Tanken är att det ska synas (diskret) vilka möjliga mätvärden som finns, för att ge referenspunkter utan att ha värden på x-axeln. Detta blir mera pedagogiskt kanske om det skulle finnas en indikation på medel/SD för någon referensgrupp.

Färgade prickar indikerar svarsfördelningen, där både storlek och färg påverkas av antalet individer som har samma mätvärde.

“Diamanten” visar medelvärdet (eftersom detta är mätvärden på intervallnivå), och de horisontala strecken visar 90% konfidensintervall. Detta eftersom det ger ett seriöst sätt att göra jämförelser. Om man har två mätningar och deras 90%CI inte överlappar är det en signifikant skillnad.

20

Code

# testing a plot of "empty" circles, filled with color gradient based on density/frequency of response
ggplot() +
  # plot possible values as hollow circles
  geom_point(
    data = df.counts20,
    aes(x = score, y = 1),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  # plot actual scores in data sample as filled circles, with gradient based on frequency(?)
  geom_point(
    data = na.omit(df.counts20), # remove missing data to remove NA from legend
    aes(x = score, y = 1, size = n * 1.5, color = nFactor),
    # size = 3,
    shape = 16
  ) +
  ### plot 95% confidence interval around the mean
  # segment makes a line
  # geom_segment(data = df.test,
  #               aes(x = mean(score) - sample90ci,
  #                   xend = mean(score) + sample90ci,
  #                   y = 0.95, yend = 0.95),
  #              color = "darkgrey"
  #            ) +
  # errorbar has brackets at the endpoints
  # geom_errorbar(
  #   data = df.test,
  #   aes(
  #     xmin = mean(score) - sample90ci,
  #     xmax = mean(score) + sample90ci,
  #     y = 0.95,
  #     width = 0.025
  #   ),
  #   color = "darkgrey"
  # ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.test20,
    aes(x = mean(score), y = 0.95),
    size = 7,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(0.8, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 1",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart diamant indikerar medelvärde.\nTomma ringar indikerar möjliga mätvärden.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank()
  ) +
  scale_y_continuous(minor_breaks = NULL) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

7

Code

# testing a plot of "empty" circles, filled with color gradient based on density/frequency of response
ggplot() +
  # plot possible values as hollow circles
  geom_point(
    data = df.counts7,
    aes(x = score, y = 1),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  # plot actual scores in data sample as filled circles, with gradient based on frequency(?)
  geom_point(
    data = na.omit(df.counts7), # remove missing data to remove NA from legend
    aes(x = score, y = 1, size = n * 1.5, color = nFactor),
    # size = 3,
    shape = 16
  ) +
  ### plot 95% confidence interval around the mean
  # segment makes a line
  # geom_segment(data = df.test,
  #               aes(x = mean(score) - sample90ci,
  #                   xend = mean(score) + sample90ci,
  #                   y = 0.95, yend = 0.95),
  #              color = "darkgrey"
  #            ) +
  # errorbar has brackets at the endpoints
  # geom_errorbar(
  #   data = df.test,
  #   aes(
  #     xmin = mean(score) - sample90ci,
  #     xmax = mean(score) + sample90ci,
  #     y = 0.95,
  #     width = 0.025
  #   ),
  #   color = "darkgrey"
  # ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.test7,
    aes(x = mean(score), y = 0.95),
    size = 7,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(0.8, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 1",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart diamant indikerar medelvärde.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank()
  ) +
  scale_y_continuous(minor_breaks = NULL) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

40

Code

# testing a plot of "empty" circles, filled with color gradient based on density/frequency of response
ggplot() +
  # plot possible values as hollow circles
  geom_point(
    data = df.counts40,
    aes(x = score, y = 1),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  # plot actual scores in data sample as filled circles, with gradient based on frequency(?)
  geom_point(
    data = na.omit(df.counts40), # remove missing data to remove NA from legend
    aes(x = score, y = 1, size = n * 1.5, color = nFactor),
    # size = 3,
    shape = 16
  ) +
  ### plot 95% confidence interval around the mean
  # segment makes a line
  # geom_segment(data = df.test,
  #               aes(x = mean(score) - sample90ci,
  #                   xend = mean(score) + sample90ci,
  #                   y = 0.95, yend = 0.95),
  #              color = "darkgrey"
  #            ) +
  # errorbar has brackets at the endpoints
  # geom_errorbar(
  #   data = df.test,
  #   aes(
  #     xmin = mean(score) - sample90ci,
  #     xmax = mean(score) + sample90ci,
  #     y = 0.95,
  #     width = 0.025
  #   ),
  #   color = "darkgrey"
  # ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.test40,
    aes(x = mean(score), y = 0.95),
    size = 7,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(0.8, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 1",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart diamant indikerar medelvärde.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank()
  ) +
  scale_y_continuous(minor_breaks = NULL) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

14.3.2 Förslag 2

Detta ser kanske något suboptimalt ut när det bara är ett index (eller en grupp/mätning) som visualiseras.

En möjlighet med stat_dots är att lägga till ett lager som visar en jämförande mätning med dotplot som går nedåt (spegelvänt runt samma linje), i en annan kulör.

20

Code

ggplot(
  df.test20,
  aes(x = score, y = 0)
) +
  geom_point(
    data = df.counts20,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  stat_dots(
    side = "top",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    color = "white",
    alpha = 1,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    fill = prevent_light_green,
    dotsize = 2
  ) +
  # plot mean
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(-0.1, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  labs(
    title = "Förslag 2",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart punkt indikerar medelvärde. Tomma cirklar indikerar möjliga mätvärden.",
    y = "",
    x = ""
  ) +
    annotate("text", x = 2.7, y = -0.13, label = "Högsta möjliga värde",
           color = "darkgrey") +
  geom_curve(x = 3.5, y = -0.14,
           xend = 4, yend = -0.03,
           color = "darkgrey",
           #curvature = -0.4,
           arrow = arrow()
           ) +
  annotate("text", x = -2.3, y = -0.13, label = "Lägsta möjliga värde",
           color = "darkgrey") +
  geom_curve(x = -2.8, y = -0.1,
           xend = -2.27, yend = -0.02, # xend = lägsta värdet i `df.counts20`
           color = "darkgrey",
           curvature = -0.4,
           arrow = arrow()
           ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

20 med CI

Code

ggplot(
  df.test20,
  aes(x = score, y = 0) # y is arbitrary to just have a baseline when only visualizing one domain/index
) +
  # plot possible response values as empty circles
  geom_point(
    data = df.counts20,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  # plot distribution of actual response values
  stat_dots(
    side = "top",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    color = "white",
    alpha = 1,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    fill = prevent_light_green,
    dotsize = 2
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    aes(
      xmin = mean(score) - medSample90ci,
      xmax = mean(score) + medSample90ci,
      y = 0,
      width = 0.025
    ),
    color = "black"
  ) +
  # plot mean
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(-0.1, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  labs(
    title = "Förslag 2",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart punkt indikerar medelvärde med 90% konfidensintervall runt. 
    Tomma cirklar indikerar möjliga mätvärden.",
    y = "",
    x = ""
  ) +
  annotate("text", x = 2.7, y = -0.13, label = "Högsta möjliga värde",
           color = "darkgrey") +
  geom_curve(x = 3.5, y = -0.14,
           xend = 4, yend = -0.03,
           color = "darkgrey",
           #curvature = -0.4,
           arrow = arrow()
           ) +
  annotate("text", x = -2.3, y = -0.13, label = "Lägsta möjliga värde",
           color = "darkgrey") +
  geom_curve(x = -2.8, y = -0.1,
           xend = -2.27, yend = -0.02, # xend = lägsta värdet i `df.counts20`
           color = "darkgrey",
           curvature = -0.4,
           arrow = arrow()
           ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

7

Code

ggplot(
  df.test7,
  aes(x = score, y = 0)
) +
  # stat_slab(
  #   side = "right", show.legend = F,
  #   scale = 0.6, # defines the height that a slab can reach
  #   position = position_dodge(width = .6), # distance between elements for dodging
  #   aes(fill_ramp = after_stat(level), fill = group),
  #   .width = c(.50, .75, 1)
  # ) +
  geom_point(
    data = df.counts7,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  stat_dots(
    side = "top",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    color = "white",
    alpha = 1,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    fill = prevent_light_green,
    dotsize = 2
  ) +
  # errorbar has brackets at the endpoints
  # geom_errorbar(
  #   aes(
  #     xmin = mean(score) - smallSample90ci,
  #     xmax = mean(score) + smallSample90ci,
  #     y = 0,
  #     width = 0.025
  #   ),
  #   color = "black"
  # ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(-0.1, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  # scale_color_viridis_d(
  #   guide = "none",
  #   begin = 0.6,
  #   aesthetics = c("fill", "color"),
  #   direction = 1
  # ) +
  labs(
    title = "Förslag 2",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart punkt indikerar medelvärde.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

40

Code

ggplot(
  df.test40,
  aes(x = score, y = 0)
) +
  # stat_slab(
  #   side = "right", show.legend = F,
  #   scale = 0.6, # defines the height that a slab can reach
  #   position = position_dodge(width = .6), # distance between elements for dodging
  #   aes(fill_ramp = after_stat(level), fill = group),
  #   .width = c(.50, .75, 1)
  # ) +
  geom_point(
    data = df.counts40,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  stat_dots(
    side = "top",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    color = "white",
    alpha = 1,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    fill = prevent_light_green,
    dotsize = 2
  ) +
  # errorbar has brackets at the endpoints
  # geom_errorbar(
  #   aes(
  #     xmin = mean(score) - smallSample90ci,
  #     xmax = mean(score) + smallSample90ci,
  #     y = 0,
  #     width = 0.025
  #   ),
  #   color = "black"
  # ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(-0.1, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  # scale_color_viridis_d(
  #   guide = "none",
  #   begin = 0.6,
  #   aesthetics = c("fill", "color"),
  #   direction = 1
  # ) +
  labs(
    title = "Förslag 2",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart punkt indikerar medelvärde.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

14.3.3 Förslag 3

20

Code

ggplot(
  df.test20,
  aes(x = score, y = 0, fill = factor(score))
) +
  # plot actual values in group
    geom_point(
    data = df.counts20,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  # plot possible values
  geom_dotplot(color = "white",
               dotsize = 1,
               binwidth = 0.2) +
   # plot mean value
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 6,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    #   ylim = c(0, 1), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  ## theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2,
    aesthetics = c("fill", "color"),
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 3",
    subtitle = "Värden längre till höger är bättre",
    #caption = "Svart diamant indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank()
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

20 med CI

Code

ggplot(
  df.test20,
  aes(x = score, y = 0, fill = factor(score))
) +
    # plot possible values
    geom_point(
    data = df.counts20,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
    # plot actual values in group
  geom_dotplot(color = "white",
               dotsize = 1,
               binwidth = 0.2) +
   # plot mean value
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 6,
    shape = 18,
    color = "black"
  ) +
    # errorbar has brackets at the endpoints
  geom_errorbar(
    aes(
      xmin = mean(score) - medSample90ci,
      xmax = mean(score) + medSample90ci,
      y = 0,
      width = 0.025
    ),
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    #   ylim = c(0, 1), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  ## theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2,
    aesthetics = c("fill", "color"),
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 3",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart diamant indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank()
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

7

Code

ggplot(
  df.test7,
  aes(x = score, y = 0, fill = factor(score))
) +
  # plot actual values in group
    geom_point(
    data = df.counts7,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  # plot possible values
  geom_dotplot(color = "white",
               dotsize = 1,
               binwidth = 0.2) +
   # plot mean value
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 6,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    #   ylim = c(0, 1), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  ## theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2,
    aesthetics = c("fill", "color"),
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 3",
    subtitle = "Värden längre till höger är bättre",
    #caption = "Svart diamant indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank()
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

40

Code

ggplot(
  df.test40,
  aes(x = score, y = 0, fill = factor(score))
) +
  # plot actual values in group
    geom_point(
    data = df.counts40,
    aes(x = score, y = -0.02, fill = NULL),
    size = 3,
    shape = 1,
    color = "grey",
    alpha = 0.7
  ) +
  # plot possible values
  geom_dotplot(color = "white",
               dotsize = 1,
               binwidth = 0.2) +
   # plot mean value
  geom_point(
    aes(
      x = mean(score),
      y = 0
    ),
    size = 6,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    #   ylim = c(0, 1), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  ## theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2,
    aesthetics = c("fill", "color"),
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 3",
    subtitle = "Värden längre till höger är bättre",
    #caption = "Svart diamant indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank()
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

14.4 Jämförelser - två mätningar samma område

14.4.1 Jämförelse 1

20

Code

# testing a plot of "empty" circles, filled with color gradient based on density/frequency of response
ggplot(
  data = df.compare20,
  aes(group = group)
) +
  # plot possible values as hollow circles
  # geom_point(data = df.countsComp,
  #            aes(x = score,
  #                y = group
  #                ),
  #            size = 3,
  #            shape = 1,
  #            color = "grey",
  #            alpha = 0.6
  # ) +
  # plot actual scores in data sample as filled circles, with gradient based on frequency(?)
  geom_point(
    data = na.omit(df.countsComp20), # remove missing data to remove NA from legend
    aes(
      x = score,
      y = group,
      size = n * 1.5,
      color = nFactor
    ),
    shape = 16
  ) +
  ## create outlines/edges for points
  # geom_point(data = na.omit(df.countsComp20), # remove missing data to remove NA from legend
  #   aes(
  #     x = score,
  #     y = group,
  #     size = n * 1.5,
  #   ),
  #   color = "white",
  #   shape = 1) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare20 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - medSample90ci,
      xmax = mean(score) + medSample90ci,
      y = group,
      width = 0.2
    ),
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  # plot mean
  geom_point(
    data = df.compare20 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
    geom_errorbar(
    data = df.compare20 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - medSample90ci2,
      xmax = mean(score) + medSample90ci2,
      y = group,
      width = 0.2
    ),
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  # plot mean
  geom_point(
    data = df.compare20 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  ### graphing properties below
  coord_cartesian(xlim = c(-3, 4)) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 1 jämförelse",
    subtitle = "Värden längre till höger är bättre",
    caption = "Grå diamant indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.title = element_blank()
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

7

Code

# testing a plot of "empty" circles, filled with color gradient based on density/frequency of response
ggplot(
  data = df.compare7,
  aes(group = group)
) +
  # plot possible values as hollow circles
  # geom_point(data = df.countsComp,
  #            aes(x = score,
  #                y = group
  #                ),
  #            size = 3,
  #            shape = 1,
  #            color = "grey",
  #            alpha = 0.6
  # ) +
  # plot actual scores in data sample as filled circles, with gradient based on frequency(?)
  geom_point(
    data = na.omit(df.countsComp7), # remove missing data to remove NA from legend
    aes(
      x = score,
      y = group,
      size = n * 1.5,
      color = nFactor
    ),
    shape = 16
  ) +
  ## create outlines/edges for points
  # geom_point(data = na.omit(df.countsComp7), # remove missing data to remove NA from legend
  #   aes(
  #     x = score,
  #     y = group,
  #     size = n * 1.5,
  #   ),
  #   color = "white",
  #   shape = 1) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare7 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - smallSample90ci,
      xmax = mean(score) + smallSample90ci,
      y = group,
      width = 0.2
    ),
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  # plot mean
  geom_point(
    data = df.compare7 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
    geom_errorbar(
    data = df.compare7 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - smallSample90ci2,
      xmax = mean(score) + smallSample90ci2,
      y = group,
      width = 0.2
    ),
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  # plot mean
  geom_point(
    data = df.compare7 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  ### graphing properties below
  coord_cartesian(xlim = c(-3, 4)) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 1 jämförelse",
    subtitle = "Värden längre till höger är bättre",
    caption = "Grå diamant indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.title = element_blank()
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

40

Code

# testing a plot of "empty" circles, filled with color gradient based on density/frequency of response
ggplot(
  data = df.compare40,
  aes(group = group)
) +
  # plot possible values as hollow circles
  # geom_point(data = df.countsComp,
  #            aes(x = score,
  #                y = group
  #                ),
  #            size = 3,
  #            shape = 1,
  #            color = "grey",
  #            alpha = 0.6
  # ) +
  # plot actual scores in data sample as filled circles, with gradient based on frequency(?)
  geom_point(
    data = na.omit(df.countsComp40), # remove missing data to remove NA from legend
    aes(
      x = score,
      y = group,
      size = n * 1.5,
      color = nFactor
    ),
    shape = 16
  ) +
  ## create outlines/edges for points
  # geom_point(data = na.omit(df.countsComp40), # remove missing data to remove NA from legend
  #   aes(
  #     x = score,
  #     y = group,
  #     size = n * 1.5,
  #   ),
  #   color = "white",
  #   shape = 1) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare40 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - largeSample90ci,
      xmax = mean(score) + largeSample90ci,
      y = group,
      width = 0.2
    ),
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  # plot mean
  geom_point(
    data = df.compare40 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
    geom_errorbar(
    data = df.compare40 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - largeSample90ci2,
      xmax = mean(score) + largeSample90ci2,
      y = group,
      width = 0.2
    ),
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  # plot mean
  geom_point(
    data = df.compare40 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "darkgrey",
    position = position_nudge(y = -0.25)
  ) +
  ### graphing properties below
  coord_cartesian(xlim = c(-3, 4)) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 1 jämförelse",
    subtitle = "Värden längre till höger är bättre",
    caption = "Grå diamant indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.title = element_blank()
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

14.4.2 Jämförelse 2

20

Code

ggplot(
  df.compare20,
  aes(x = score, y = group, fill = group)
) +
  stat_dots(
    side = "top",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    color = "white",
    alpha = 0.8,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2
  ) +
  # plot mean (point) and confidence interval (error bar) to make it easier to compare over time
  geom_errorbar(data = df.compare20 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - medSample90ci,
      xmax = mean(score) + medSample90ci,
      y = group,
      width = 0.1
    ),
    color = "black"
  ) +
  geom_point(data = df.compare20 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
    geom_errorbar(data = df.compare20 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - medSample90ci2,
      xmax = mean(score) + medSample90ci2,
      y = group,
      width = 0.1
    ),
    color = "black"
  ) +
  geom_point(data = df.compare20 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    # ylim = c(0, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_manual(
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color"),
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 2",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart punkt indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from x axis
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

7

Code

ggplot(
  df.compare7,
  aes(x = score, y = group, fill = group)
) +
  # stat_slab(
  #   side = "right", show.legend = F,
  #   scale = 0.6, # defines the height that a slab can reach
  #   position = position_dodge(width = .6), # distance between elements for dodging
  #   aes(fill_ramp = after_stat(level), fill = group),
  #   .width = c(.50, .75, 1)
  # ) +
  stat_dots(
    side = "top",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    color = "white",
    alpha = 0.8,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(data = df.compare7 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - smallSample90ci,
      xmax = mean(score) + smallSample90ci,
      y = group,
      width = 0.1
    ),
    color = "black"
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(data = df.compare7 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
    geom_errorbar(data = df.compare7 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - smallSample90ci2,
      xmax = mean(score) + smallSample90ci2,
      y = group,
      width = 0.1
    ),
    color = "black"
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(data = df.compare7 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    # ylim = c(0, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_manual(
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color"),
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 2",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart punkt indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

40

Code

ggplot(
  df.compare40,
  aes(x = score, y = group, fill = group)
) +
  # stat_slab(
  #   side = "right", show.legend = F,
  #   scale = 0.6, # defines the height that a slab can reach
  #   position = position_dodge(width = .6), # distance between elements for dodging
  #   aes(fill_ramp = after_stat(level), fill = group),
  #   .width = c(.50, .75, 1)
  # ) +
  stat_dots(
    side = "top",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    color = "white",
    alpha = 0.8,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(data = df.compare40 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - largeSample90ci,
      xmax = mean(score) + largeSample90ci,
      y = group,
      width = 0.1
    ),
    color = "black"
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(data = df.compare40 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
    geom_errorbar(data = df.compare40 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - largeSample90ci2,
      xmax = mean(score) + largeSample90ci2,
      y = group,
      width = 0.1
    ),
    color = "black"
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(data = df.compare40 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = group
    ),
    size = 5,
    shape = 18,
    color = "black"
  ) +
  ### graphing properties below
  coord_cartesian(
    # ylim = c(0, 1.2), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_manual(
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color"),
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 2",
    subtitle = "Värden längre till höger är bättre",
    caption = "Svart punkt indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

14.4.3 Jämförelse 3

20

Code

nudge <- 0.3

ggplot(
  df.compare20,
  aes(x = score, y = 0.5, fill = group)
) +
  # stat_slab(
  #   side = "right", show.legend = F,
  #   scale = 0.6, # defines the height that a slab can reach
  #   position = position_dodge(width = .6), # distance between elements for dodging
  #   aes(fill_ramp = after_stat(level), fill = group),
  #   .width = c(.50, .75, 1)
  # ) +
  stat_dots(
    data = df.compare20 %>% filter(group == "Mättillfälle 1"),
    side = "top",
    scale = 0.3,
    show.legend = T,
    position = position_dodge(width = .9),
    alpha = 0.66,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2,
    color = "white"
  ) +
  stat_dots(
    data = df.compare20 %>% filter(group == "Mättillfälle 2"),
    side = "bottom",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    alpha = 0.66,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2,
    color = "white"
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare20 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - medSample90ci2,
      xmax = mean(score) + medSample90ci2,
      y = 0.5,
      width = 0.025,
      color = group
    ),
    position = position_nudge(y = -nudge)
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.compare20 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = 0.5,
      color = group
    ),
    size = 5,
    shape = 18,
    position = position_nudge(y = -nudge),
    alpha = 0.7
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare20 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - medSample90ci,
      xmax = mean(score) + medSample90ci,
      y = 0.5,
      width = 0.025,
      color = group
    ),
    position = position_nudge(y = nudge)
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.compare20 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = 0.5,
      color = group
    ),
    size = 5,
    shape = 18,
    position = position_nudge(y = nudge),
    alpha = 0.7
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(0, 1), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_manual("Grupp",
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 3",
    subtitle = "Värden längre till höger är bättre. \nPrickar motsvarar personer.",
    caption = "Svart punkt indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(), # remove text from both axes
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

7

Code

nudge <- 0.3

ggplot(
  df.compare7,
  aes(x = score, y = 0.5, fill = group)
) +
  # stat_slab(
  #   side = "right", show.legend = F,
  #   scale = 0.6, # defines the height that a slab can reach
  #   position = position_dodge(width = .6), # distance between elements for dodging
  #   aes(fill_ramp = after_stat(level), fill = group),
  #   .width = c(.50, .75, 1)
  # ) +
  stat_dots(
    data = df.compare7 %>% filter(group == "Mättillfälle 1"),
    side = "top",
    scale = 0.3,
    show.legend = T,
    position = position_dodge(width = .9),
    alpha = 0.66,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2,
    color = "white"
  ) +
  stat_dots(
    data = df.compare7 %>% filter(group == "Mättillfälle 2"),
    side = "bottom",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    alpha = 0.66,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2,
    color = "white"
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare7 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - smallSample90ci2,
      xmax = mean(score) + smallSample90ci2,
      y = 0.5,
      width = 0.025,
      color = group
    ),
    position = position_nudge(y = -nudge)
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.compare7 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = 0.5,
      color = group
    ),
    size = 5,
    shape = 18,
    position = position_nudge(y = -nudge),
    alpha = 0.7
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare7 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - smallSample90ci,
      xmax = mean(score) + smallSample90ci,
      y = 0.5,
      width = 0.025,
      color = group
    ),
    position = position_nudge(y = nudge)
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.compare7 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = 0.5,
      color = group
    ),
    size = 5,
    shape = 18,
    position = position_nudge(y = nudge),
    alpha = 0.7
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(0, 1), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_manual("Grupp",
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 3",
    subtitle = "Värden längre till höger är bättre. \nPrickar motsvarar personer.",
    caption = "Svart punkt indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(), # remove text from both axes
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

40

Code

nudge <- 0.3

ggplot(
  df.compare40,
  aes(x = score, y = 0.5, fill = group)
) +
  # stat_slab(
  #   side = "right", show.legend = F,
  #   scale = 0.6, # defines the height that a slab can reach
  #   position = position_dodge(width = .6), # distance between elements for dodging
  #   aes(fill_ramp = after_stat(level), fill = group),
  #   .width = c(.50, .75, 1)
  # ) +
  stat_dots(
    data = df.compare40 %>% filter(group == "Mättillfälle 1"),
    side = "top",
    scale = 0.3,
    show.legend = T,
    position = position_dodge(width = .9),
    alpha = 0.66,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2,
    color = "white"
  ) +
  stat_dots(
    data = df.compare40 %>% filter(group == "Mättillfälle 2"),
    side = "bottom",
    scale = 0.3,
    show.legend = F,
    position = position_dodge(width = .9),
    alpha = 0.66,
    # verbose = T,
    binwidth = 0.1, # this and the one below may need to be calculated from data instead
    dotsize = 2,
    color = "white"
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare40 %>% filter(group == "Mättillfälle 2"),
    aes(
      xmin = mean(score) - largeSample90ci2,
      xmax = mean(score) + largeSample90ci2,
      y = 0.5,
      width = 0.025,
      color = group
    ),
    position = position_nudge(y = -nudge)
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.compare40 %>% filter(group == "Mättillfälle 2"),
    aes(
      x = mean(score),
      y = 0.5,
      color = group
    ),
    size = 5,
    shape = 18,
    position = position_nudge(y = -nudge),
    alpha = 0.7
  ) +
  # errorbar has brackets at the endpoints
  geom_errorbar(
    data = df.compare40 %>% filter(group == "Mättillfälle 1"),
    aes(
      xmin = mean(score) - largeSample90ci,
      xmax = mean(score) + largeSample90ci,
      y = 0.5,
      width = 0.025,
      color = group
    ),
    position = position_nudge(y = nudge)
  ) +
  # plot mean/median and some sort of distribution (SD/SE?) to make it easier to compare over time
  geom_point(
    data = df.compare40 %>% filter(group == "Mättillfälle 1"),
    aes(
      x = mean(score),
      y = 0.5,
      color = group
    ),
    size = 5,
    shape = 18,
    position = position_nudge(y = nudge),
    alpha = 0.7
  ) +
  ### graphing properties below
  coord_cartesian(
    ylim = c(0, 1), # limits for x and y axis
    xlim = c(-3, 4)
  ) +
  scale_size_continuous(
    range = c(5, 10), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_color_manual("Grupp",
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Förslag 3",
    subtitle = "Värden längre till höger är bättre. \nPrickar motsvarar personer.",
    caption = "Svart punkt indikerar medelvärde med 90% konfidensintervall runt.",
    y = "",
    x = ""
  ) +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(), # remove text from both axes
    axis.title = element_blank(),
  ) +
  guides(color = guide_legend(override.aes = list(size = 7))) # make points in legend bigger

14.5 Visualisering av frågor/items

Code

# create random sample dataset
df.plot7 <- df %>%
  slice_sample(n = sampleSmall) %>%
  select(all_of(itemlabels$itemnr), score) %>%
  pivot_longer(!score) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 1")

# enable comparisons by adding another random group
df.plot7b <- df %>%
  slice_sample(n = sampleSmall) %>%
  select(all_of(itemlabels$itemnr), score) %>%
  pivot_longer(!score) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 2")

df.plotComp7 <- rbind(df.plot7,df.plot7b)

#####

# create random sample dataset
df.plot20 <- df %>%
  slice_sample(n = sampleMed) %>%
  select(all_of(itemlabels$itemnr), score) %>%
  pivot_longer(!score) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 1")

# enable comparisons by adding another random group
df.plot20b <- df %>%
  slice_sample(n = sampleMed) %>%
  select(all_of(itemlabels$itemnr), score) %>%
  pivot_longer(!score) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 2")

df.plotComp20 <- rbind(df.plot20,df.plot20b)

#####

# create random sample dataset
df.plot40 <- df %>%
  slice_sample(n = sampleLarge) %>%
  select(all_of(itemlabels$itemnr), score) %>%
  pivot_longer(!score) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 1")

# enable comparisons by adding another random group
df.plot40b <- df %>%
  slice_sample(n = sampleLarge) %>%
  select(all_of(itemlabels$itemnr), score) %>%
  pivot_longer(!score) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 2")

df.plotComp40 <- rbind(df.plot40,df.plot40b)

Code

# function to get mode values (typvärden)
Mode <- function(x) {
  ux <- unique(x)
  ux[which.max(tabulate(match(x, ux)))]
}

# prepare dataframe to store values
modeResponses <- itemlabels

df.modes <- df.plot20 %>% 
  # create numeric responses where 1 = "Aldrig, and 6 = "Alltid".
  mutate(svarNum = as.integer(fct_rev(svarskategori))) %>% 
  add_column(id = rep(1:20, each=6)) %>% 
  select(itemnr,svarNum,id) %>% 
  pivot_wider(names_from = "itemnr",
              values_from = "svarNum",
              id_cols = "id")

modes <- c()
for (i in modeResponses$itemnr) {
  mode1 <- Mode(df.modes[[i]])
  modes <- c(modes,mode1)
}
modeResponses$modes <- modes

# get median values (can be .5 when we have an even N)
medians <- c()
for (i in modeResponses$itemnr) {
  med1 <- median(df.modes[[i]])
  medians <- c(medians,med1)
}

modeResponses$medians <- medians

# get IQR, interquartile range
iqrs <- c()
for (i in modeResponses$itemnr) {
  iqr1 <- IQR(df.modes[[i]])
  iqrs <- c(iqrs,iqr1)
}
modeResponses$iqrs <- iqrs

# get median values (can be .5 when we have an even N) and upper/lower IQR
med_hilow <- matrix(nrow = 0, ncol = 3)
for (i in modeResponses$itemnr) {
  med_hilow <- rbind(med_hilow,median_hilow(df.modes[[i]], conf.int=.5))
}

modeResponses <- cbind(modeResponses,med_hilow)

modeResponses <- modeResponses %>% 
  mutate(modeCats = factor(modes, levels = c(1:6),
                           labels = svarskategorier)
         )

Code

# prepare dataframe to store values
modeResponses7 <- itemlabels

df.modes7 <- df.plot7 %>% 
  # create numeric responses where 1 = "Aldrig, and 6 = "Alltid".
  mutate(svarNum = as.integer(fct_rev(svarskategori))) %>% 
  add_column(id = rep(1:7, each=6)) %>% 
  select(itemnr,svarNum,id) %>% 
  pivot_wider(names_from = "itemnr",
              values_from = "svarNum",
              id_cols = "id")

modes <- c()
for (i in modeResponses$itemnr) {
  mode1 <- Mode(df.modes7[[i]])
  modes <- c(modes,mode1)
}
modeResponses7$modes <- modes

# get median values (can be .5 when we have an even N)
medians <- c()
for (i in modeResponses$itemnr) {
  med1 <- median(df.modes7[[i]])
  medians <- c(medians,med1)
}

modeResponses7$medians <- medians

# get IQR, interquartile range
iqrs <- c()
for (i in modeResponses$itemnr) {
  iqr1 <- IQR(df.modes7[[i]])
  iqrs <- c(iqrs,iqr1)
}
modeResponses7$iqrs <- iqrs

# get median values (can be .5 when we have an even N) and upper/lower IQR
med_hilow7 <- matrix(nrow = 0, ncol = 3)
for (i in modeResponses$itemnr) {
  med_hilow7 <- rbind(med_hilow7,median_hilow(df.modes7[[i]], conf.int=.5))
}

modeResponses7 <- cbind(modeResponses7,med_hilow7)

modeResponses7 <- modeResponses7 %>% 
  mutate(modeCats = factor(modes, levels = c(1:6),
                           labels = svarskategorier)
         )

Code

# prepare dataframe to store values
modeResponses40 <- itemlabels

df.modes40 <- df.plot40 %>% 
  # create numeric responses where 1 = "Aldrig, and 6 = "Alltid".
  mutate(svarNum = as.integer(fct_rev(svarskategori))) %>% 
  add_column(id = rep(1:40, each=6)) %>% 
  select(itemnr,svarNum,id) %>% 
  pivot_wider(names_from = "itemnr",
              values_from = "svarNum",
              id_cols = "id")

modes <- c()
for (i in modeResponses$itemnr) {
  mode1 <- Mode(df.modes40[[i]])
  modes <- c(modes,mode1)
}
modeResponses40$modes <- modes

# get median values (can be .5 when we have an even N)
medians <- c()
for (i in modeResponses$itemnr) {
  med1 <- median(df.modes40[[i]])
  medians <- c(medians,med1)
}

modeResponses40$medians <- medians

# get IQR, interquartile range
iqrs <- c()
for (i in modeResponses$itemnr) {
  iqr1 <- IQR(df.modes40[[i]])
  iqrs <- c(iqrs,iqr1)
}
modeResponses40$iqrs <- iqrs

# get median values (can be .5 when we have an even N) and upper/lower IQR
med_hilow40 <- matrix(nrow = 0, ncol = 3)
for (i in modeResponses$itemnr) {
  med_hilow40 <- rbind(med_hilow40,median_hilow(df.modes40[[i]], conf.int=.5))
}

modeResponses40 <- cbind(modeResponses40,med_hilow40)

modeResponses40 <- modeResponses40 %>% 
  mutate(modeCats = factor(modes, levels = c(1:6),
                           labels = svarskategorier)
         )

14.5.1 geom_point

20

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "none") +
  scale_color_viridis_d("",
    begin = 0.2,
    end = 0.8,
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  #guides(color = guide_legend(override.aes = list(size = 7))) + # make points in legend bigger
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 med text

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  geom_text(aes(x = svarskategori, y = item, label = n),
            color = "white") +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "none") +
  scale_color_viridis_d("",
    begin = 0.2,
    end = 0.8,
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  #guides(color = guide_legend(override.aes = list(size = 7))) + # make points in legend bigger
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 text+blå kulör

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    # size = 3,
    shape = 16
  ) +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  geom_text(aes(x = svarskategori, y = item, label = n),
            color = "white") +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_brewer(
  #   #type = "seq",
  #   #palette = "Greens",
   guide = "none"
  ) +
  #scale_color_manual(values = my.cols) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 blå utan text

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16
  ) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_brewer(
   guide = "none"
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

7

Code

df.plot7 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  scale_size_continuous(
    range = c(7, 18), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "none") +
  scale_color_viridis_d("",
    begin = 0.2,
    end = 0.8,
    guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  #guides(color = guide_legend(override.aes = list(size = 7))) + # make points in legend bigger
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

7 blå

Code

df.plot7 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16
  ) +
  scale_size_continuous(
    range = c(7, 18), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "none") +
  scale_color_brewer(
   guide = "none"
  ) +
  # scale_color_manual(values = PREVENTpalette1) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  #guides(color = guide_legend(override.aes = list(size = 7))) + # make points in legend bigger
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40

Code

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  scale_size_continuous(
    range = c(7, 21), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "none") +
  scale_color_viridis_d("",
    begin = 0.2,
    end = 0.8,
    guide = "none"
  ) +
  #scale_color_manual(values = plot40colors) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  #guides(color = guide_legend(override.aes = list(size = 7))) + # make points in legend bigger
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40 blå

Code

# scale_color_brewer() only has 9 colors in its palette, we need more for larger samples
# Let's calculate the needed number first

numColors <- df.plot40 %>%
  dplyr::count(item, svarskategori) %>% 
  distinct(n) %>% 
  rownames_to_column("categories") %>% 
  mutate(categories = as.numeric(categories)) %>% 
  select(categories) %>% 
  tail(1) %>% 
  pull()

plot40colors <- colorRampPalette(c("#F7FBFF","#08306B"))(numColors)

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16
  ) +
  scale_size_continuous(
    range = c(7, 18), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "none") +
  # scale_color_brewer(
  #  guide = "none"
  # ) +
  scale_color_manual(values = plot40colors) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  #guides(color = guide_legend(override.aes = list(size = 7))) + # make points in legend bigger
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

14.5.2 Typvärden/medianvärden 20

20 median + bollar

Medianvärdet är det som ligger i mitten när alla värden i gruppen ställs på rad från lågt till högt. Vi börjar med att titta på föregående visualisering “i bakgrunden” och lägga över median-värdet.

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16,
    alpha = 0.6
  ) +
  geom_point(data = modeResponses,
             aes(x = medians,
                 y = item),
             color = "black",
             size = 7,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_brewer(
   guide = "none"
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar + medianvärden",
    caption = "Medianvärdet kan hamna mellan två svarskategorier när det är ett jämnt antal respondenter.",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 typvärde + bollar

Typvärdet är det vanligast förekommande svaret på en fråga.

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16,
    alpha = 0.6
  ) +
  geom_point(data = modeResponses,
             aes(x = modes,
                 y = item),
             color = "black",
             size = 7,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_brewer(
   guide = "none"
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar + typvärden",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 enbart typvärde

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5),
    shape = 16,
    alpha = 0
  ) +
  geom_point(data = modeResponses,
             aes(x = modes,
                 y = item,
                 color = modeCats),
             #color = "black",
             size = 12,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d(
   guide = "none",
   begin = 0.4,
   end = 0.8
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Typvärden",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 median + spridning

Vi använder interkvartilt omfång (interquartile range, IQR), vilket är avståndet mellan 25:e och 75:e percentilen. Med mindre sampel är dock detta mindre användbart. Vi tittar på sampelstorlek 20, med “bollformade” figurer i bakgrunden för att visa den faktiska spridningen samtidigt.

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5),
    shape = 16,
    alpha = 0.05
  ) +
  geom_segment(data = modeResponses,
             aes(
               y = item, yend = item,
               x = ymin,
               xend = ymax
               ),
             lineend = "round",
             linewidth = 2,
             color = "black",
             alpha = 0.7
             ) +
  geom_point(data = modeResponses,
             aes(x = medians,
                 y = item,
                 color = modeCats),
             #color = "black",
             size = 12,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d(
   guide = "none",
   begin = 0.4,
   end = 0.8
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Median och spridning",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 median+spridning utan bollar

Code

df.plot20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5),
    shape = 16,
    alpha = 0
  ) +
  geom_segment(data = modeResponses,
             aes(
               y = item, yend = item,
               x = ymin,
               xend = ymax
               ),
             lineend = "round",
             linewidth = 2,
             color = "black",
             alpha = 0.7
             ) +
  geom_point(data = modeResponses,
             aes(x = medians,
                 y = item,
                 color = modeCats),
             #color = "black",
             size = 12,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d(
   guide = "none",
   begin = 0.4,
   end = 0.8
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Median och spridning",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

14.5.3 Typ/median 40

40 median + bollar

Vi börjar med att titta på föregående visualisering “i bakgrunden” och lägga över median-värdet.

Medianvärdet är det som ligger i mitten när alla värden i gruppen ställs på rad från lågt till högt.

Code

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16,
    alpha = 0.6
  ) +
  geom_point(data = modeResponses40,
             aes(x = modes,
                 y = item),
             color = "black",
             size = 7,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_manual(values = plot40colors,
                     guide = "none") +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar + medianvärden",
    caption = "Medianvärdet kan hamna mellan två svarskategorier när det är ett jämnt antal respondenter.",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40 typvärde + bollar

Code

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16,
    alpha = 0.6
  ) +
  geom_point(data = modeResponses40,
             aes(x = modes,
                 y = item),
             color = "black",
             size = 7,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_manual(values = plot40colors,
                     guide = "none") +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar + typvärden",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40 enbart typvärde

Code

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5),
    shape = 16,
    alpha = 0
  ) +
  geom_point(data = modeResponses40,
             aes(x = modes,
                 y = item,
                 color = modeCats),
             #color = "black",
             size = 12,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d(
   guide = "none",
   begin = 0.4,
   end = 0.8
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Typvärden",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40 median + spridning

Vi använder interkvartilt omfång (interquartile range, IQR), vilket är avståndet mellan 25:e och 75:e percentilen. Med mindre sampel är dock detta mindre användbart. Vi tittar på sampelstorlek 20, med “bollformade” figurer i bakgrunden för att visa den faktiska spridningen samtidigt.

Code

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5),
    shape = 16,
    alpha = 0.05
  ) +
  geom_segment(data = modeResponses40,
             aes(
               y = item, yend = item,
               x = ymin,
               xend = ymax
               ),
             lineend = "round",
             linewidth = 2,
             color = "black",
             alpha = 0.7
             ) +
  geom_point(data = modeResponses40,
             aes(x = medians,
                 y = item,
                 color = modeCats),
             #color = "black",
             size = 12,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d(
   guide = "none",
   begin = 0.4,
   end = 0.8
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Median och spridning",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40 median+spridning

Code

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5),
    shape = 16,
    alpha = 0
  ) +
  geom_segment(data = modeResponses40,
             aes(
               y = item, yend = item,
               x = ymin,
               xend = ymax
               ),
             lineend = "round",
             linewidth = 2,
             color = "black",
             alpha = 0.7
             ) +
  geom_point(data = modeResponses40,
             aes(x = medians,
                 y = item,
                 color = modeCats),
             #color = "black",
             size = 12,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d(
   guide = "none",
   begin = 0.4,
   end = 0.8
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Median och spridning",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

Code

df.modes40 %>% 
  select(!id) %>% 
  pivot_longer(everything(), names_to = "itemnr", values_to = "svarNum")

# A tibble: 240 × 2
   itemnr svarNum
   <chr>    <int>
 1 abk4         5
 2 abk5         3
 3 abk6         3
 4 å2           2
 5 å4           3
 6 å5           2
 7 abk4         6
 8 abk5         3
 9 abk6         3
10 å2           4
# ℹ 230 more rows

Code

df.plot40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  mutate(svarNum = as.numeric(svarskategori)) %>% 
  ggplot(aes(x = svarskategori, 
             y = item)) +
  geom_point(aes(size = n * 1.5),
    shape = 16,
    alpha = 0
  ) +
  geom_point(data = modeResponses40,
             aes(x = medians,
                 y = item,
                 color = modeCats),
             #color = "black",
             size = 12,
             shape = 18) +
  stat_pointinterval(point_interval = median_qi) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_viridis_d(
   guide = "none",
   begin = 0.4,
   end = 0.8
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Median och spridning",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

14.5.4 Typ/median 7

7 median + bollar

Det blir tyvärr inte bra när vi har såpass få deltagare i samplet. Median är något mindre dåligt, men typvärdet fungerar inte med så små sampel. Men även median känns inte rättvisade, vilket framgår i figuren nedan.

Code

df.plot7 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = nFactor),
    shape = 16,
    alpha = 0.6
  ) +
  geom_point(data = modeResponses7,
             aes(x = medians,
                 y = item),
             color = "black",
             size = 7,
             shape = 18) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_brewer(
   guide = "none"
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar + medianvärden",
    caption = "Medianvärdet kan hamna mellan två svarskategorier när det är ett jämnt antal respondenter.",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

14.5.5 stat_dots

denna borde också kunna användas för pre/post, genom att ha två layers med stat_dots för de olika mättillfällena? Och använda position_nudge för att lägga dem ovan/under y-linjen.

20

Code

df.plot20 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      color = svarskategori, fill = svarskategori
    )
  ) +
  stat_dots() +
  scale_color_viridis_d(
    begin = 0.2,
    aesthetics = c("fill", "color"),
    guide = "none",
    direction = 1
  ) +
  theme_minimal() +
  theme_prevent() +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8)) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  )

7

Code

df.plot7 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      color = svarskategori, fill = svarskategori
    )
  ) +
  stat_dots() +
  scale_color_viridis_d(
    begin = 0.2,
    aesthetics = c("fill", "color"),
    guide = "none",
    direction = 1
  ) +
  theme_minimal() +
  theme_prevent() +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8)) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  )

40

Code

df.plot40 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      color = svarskategori, fill = svarskategori
    )
  ) +
  stat_dots() +
  scale_color_viridis_d(
    begin = 0.2,
    aesthetics = c("fill", "color"),
    guide = "none",
    direction = 1
  ) +
  theme_minimal() +
  theme_prevent() +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8)) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  )

14.5.6 ridgeline_gradient

20

Code

df.plot20 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>%
  group_by(item) %>% 
  count(svarskategori) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      height = n, group = item,
      fill = svarskategori
    )) +
  geom_ridgeline_gradient(scale = 8/nrow(df.plot20)) +
  scale_x_discrete() +
  scale_fill_viridis_d('Svarskategori', 
                       direction = 1,
                       begin = 0.4, 
                       guide = "none") +
  theme_minimal() +
  theme_prevent() +
  xlab("") +
  ylab("") +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  labs(title = "Ridgeline",
       subtitle = glue("{nrow(df.plot20)/6} deltagare"),
       caption = "Siffror indikerar antal svar per svarskategori") + # dela på antalet frågor i indexet
  geom_text(aes(label = n), color = "orange", 
            nudge_y = -0.1, nudge_x = 0) +
  theme()

Code

# ridgeline med en kulör bara

7

Code

df.plot7 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>%
  group_by(item) %>% 
  count(svarskategori) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      height = n, group = item,
      fill = svarskategori
    )) +
  geom_ridgeline_gradient(scale = 8/nrow(df.plot7)) +
  scale_x_discrete() +
  scale_fill_viridis_d('Svarskategori', 
                       direction = 1,
                       begin = 0.4, 
                       guide = "none") +
  theme_minimal() +
  theme_prevent() +
  xlab("") +
  ylab("") +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  labs(title = "Ridgeline",
       subtitle = glue("{nrow(df.plot7)/6} deltagare")) +
  geom_text(aes(label = n), color = "orange", 
            nudge_y = -0.1, nudge_x = 0)

40

Code

df.plot40 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>%
  group_by(item) %>% 
  count(svarskategori) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      height = n, group = item,
      fill = svarskategori
    )) +
  geom_ridgeline_gradient(scale = 8/nrow(df.plot40)) +
  scale_x_discrete() +
  scale_fill_viridis_d('Svarskategori', 
                       direction = 1,
                       begin = 0.4, 
                       guide = "none") +
  theme_minimal() +
  theme_prevent() +
  xlab("") +
  ylab("") +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  labs(title = "Ridgeline",
       subtitle = glue("{nrow(df.plot40)/6} deltagare")) +
  geom_text(aes(label = n), color = "orange", 
            nudge_y = -0.1, nudge_x = 0)

14.5.7 density

20 enfärgad

Code

densityCount20 <- df.plot20 %>%
  #mutate(svarskategori = fct_rev(svarskategori)) %>%
  mutate(item = fct_rev(item)) %>% 
  group_by(item) %>% 
  count(svarskategori) %>% 
  mutate(svar = as.numeric(svarskategori))

df.plot20 %>%
  mutate(svar = as.numeric(svarskategori)) %>% 
  mutate(item = factor(item)) %>% 
  #mutate(item = fct_rev(item)) %>% 
  ggplot(
    aes(
      x = svar
        )
      ) +
  geom_density(alpha = 0.8, 
               stat = "count",
               fill = prevent_light_blue, 
               color = prevent_blue
               ) +
  geom_text(data = densityCount20,
            aes(y = 0, label = n), color = prevent_dark_blue, alpha = 1, 
            nudge_y = 0, nudge_x = 0) +
  coord_cartesian(clip = "off") +
  facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(30))
             ) +
  scale_x_continuous(labels = svarskategorier,
                     #limits = c(1:6), 
                     breaks = c(1:6)
    ) +
  scale_y_continuous(position="right", 
                     minor_breaks = NULL,
                     label = scales::label_comma(accuracy = 1)) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0,
                                                 hjust = 1),
                legend.position = "none"
                                ) +
  theme(axis.title.y = element_text(margin = margin(l = 15))) +
  xlab("") +
  ylab("Antal svar") +
  labs(title = "Density count",
       subtitle = "20 deltagare",
        caption = "Siffror indikerar antal svar per svarskategori")

14.5.8 geom_col + facet_wrap

14.5.8.1 20

Code

df.plot20 %>%
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  ggplot(
    aes(
      x = svarskategori, y = n, fill = svarskategori
    )
  ) +
  geom_col() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2,
    aesthetics = c("fill", "color")
  ) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "none") +  
  xlab("") +
  ylab("") +
  scale_y_continuous(position="right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL)

14.5.8.2 7

Code

df.plot7 %>%
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  ggplot(
    aes(
      x = svarskategori, y = n, fill = svarskategori
    )
  ) +
  geom_col() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2,
    aesthetics = c("fill", "color")
  ) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "none") +  
  xlab("") +
  ylab("") +
  scale_y_continuous(position="right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL)

14.5.8.3 40

Code

df.plot40 %>%
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  ggplot(
    aes(
      x = svarskategori, y = n, fill = svarskategori
    )
  ) +
  geom_col() +
  scale_color_viridis_d("Antal svar",
    begin = 0.2,
    aesthetics = c("fill", "color")
  ) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "none") +  
  xlab("") +
  ylab("") +
  scale_y_continuous(position="right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL)

14.5.9 Kränkning/särbehandling

Denna får en egen visualisering/kulörpalett.

Code

# read data again for negative acts questions, "krbet"
df.krbet <- read.spss(spssDatafil, to.data.frame = TRUE) %>% 
  select(starts_with("q0012")) %>% 
  na.omit()
# krbet itemlabels
krbet.itemlabels <- read_excel("../data/Itemlabels.xlsx") %>% 
  filter(str_detect(itemnr, pattern = "kb")) %>% 
  select(!Dimension)

names(df.krbet) <- krbet.itemlabels$itemnr

df.plot.krbet20 <- df.krbet %>%
  slice_sample(n = 20) %>%
  pivot_longer(everything()) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(krbet.itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 1")

df.plot.krbet20b <- df.krbet %>%
  slice_sample(n = 20) %>%
  pivot_longer(everything()) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(krbet.itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 2")

df.krbetComp <- rbind(df.plot.krbet20,df.plot.krbet20b)

df.plot.krbet7 <- df.krbet %>%
  slice_sample(n = 7) %>%
  pivot_longer(everything()) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(krbet.itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 1")

df.plot.krbet40 <- df.krbet %>%
  slice_sample(n = 40) %>%
  pivot_longer(everything()) %>% # we need long format for ggplot
  rename(
    itemnr = name,
    svarskategori = value
  ) %>%
  left_join(krbet.itemlabels, by = "itemnr") %>%  # get item descriptions as a variable in the df
  add_column(group = "Mättillfälle 1")

krbet.svarskategorier <- c("Aldrig","Det har hänt","Varje månad","Varje vecka","Dagligen")

20

Code

df.plot.krbet20 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  #mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  geom_text(aes(x = svarskategori, y = item, label = n),
            color = "white") +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  # scale_color_viridis_d("Antal svar",
  #   begin = 0,
  #   option = "plasma",
  #   direction = -1
  # ) +
  scale_color_manual(values = c("#008332","#FBB900","#BE5014","#B01200","#B01200"),
                     guide = "none") +
  labs(
    title = "Kränkande beteenden",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(breaks = krbet.svarskategorier,
                   limits = krbet.svarskategorier)

7

Code

df.plot.krbet7 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  #mutate(svarskategori = fct_rev(svarskategori)) %>% 
  filter(n > 0) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  geom_text(aes(x = svarskategori, y = item, label = n),
            color = "white") +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  # scale_color_viridis_d("Antal svar",
  #   begin = 0,
  #   option = "plasma",
  #   direction = -1
  # ) +
  scale_color_manual(values = c("#008332","#FBB900","#BE5014","#B01200","#B01200"),
                     guide = "none",
                     drop = FALSE) +
  labs(
    title = "Kränkande beteenden",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(breaks = krbet.svarskategorier,
                   limits = krbet.svarskategorier)

40

Code

df.plot.krbet40 %>%
  dplyr::count(item, svarskategori) %>%
  mutate(nFactor = factor(n)) %>%
  #mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot() +
  geom_point(aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  geom_text(aes(x = svarskategori, y = item, label = n),
            color = "white") +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  # scale_color_viridis_d("Antal svar",
  #   begin = 0,
  #   option = "plasma",
  #   direction = -1
  # ) +
  scale_color_manual(values = c("#008332","#FBB900","#BE5014","#B01200","#B01200"),
                     guide = "none") +
  labs(
    title = "Kränkande beteenden",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(breaks = krbet.svarskategorier,
                   limits = krbet.svarskategorier)

20 jämförelse

Code

plot.krbetComp <- df.krbetComp %>%
  dplyr::count(group, item, svarskategori) %>%
  mutate(nFactor = factor(n))
# mutate(svarskategori = fct_rev(svarskategori)) %>%
ggplot() +
  geom_point(
    data = plot.krbetComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  geom_text(
    data = plot.krbetComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, label = n),
    color = "white"
  ) +
  geom_point(
    data = plot.krbetComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    alpha = 0.3,
    shape = 16,
    position = position_nudge(x = 0.35, y = 0)
  ) +
  geom_text(
    data = plot.krbetComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, label = n),
    color = "white",
    alpha = 0.66,
    position = position_nudge(x = 0.35, y = 0)
  ) +
  ### graphing properties below
  # coord_cartesian(ylim = c(0.8,1.2), # limits for x and y axis
  #                xlim = c(-3,4)) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none"
  ) + # remove legend for size aesthetic
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent() +
  scale_color_manual(
    values = c("#008332", "#FBB900", "#BE5014", "#B01200", "#B01200"),
    guide = "none"
  ) +
  labs(
    title = "Kränkande beteenden",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(
    breaks = krbet.svarskategorier,
    limits = krbet.svarskategorier
  )

14.6 Jämförelse på item-nivå

14.6.1 columns

20

Code

df.plotComp20 %>%
  group_by(group) %>% 
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = n, fill = group
    )
  ) +
  geom_col(position = "dodge") +
  scale_color_manual('',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             #scales = "free"
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "top") +  
  xlab("") +
  ylab("") +
  scale_y_continuous(position = "right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

7

Code

df.plotComp7 %>%
  group_by(group) %>% 
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = n, fill = group
    )
  ) +
  geom_col(position = "dodge") +
  scale_color_manual('',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             #scales = "free"
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "top") +  
  xlab("") +
  ylab("") +
  scale_y_continuous(position = "right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40

Code

df.plotComp40 %>%
  group_by(group) %>% 
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = n, fill = group
    )
  ) +
  geom_col(position = "dodge") +
  scale_color_manual('',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             #scales = "free"
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "top") +  
  xlab("") +
  ylab("") +
  scale_y_continuous(position = "right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

14.6.2 density count

20

Code

df.plotComp20 %>%
  mutate(svar = as.numeric(svarskategori)) %>% 
  mutate(item = factor(item)) %>% 
  #mutate(item = fct_rev(item)) %>% 
  ggplot(
    aes(
      x = svar, fill = group, color = group
    )) +
  geom_density(alpha = 0.6, stat = "count") +
  facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             ) +
  scale_x_continuous(limits = svarskategorier, 
                     breaks = svarskategorier,
                     labels = svarskategorier
    ) +
  scale_y_continuous(position="right", 
                     minor_breaks = NULL,
                     label = scales::label_comma(accuracy = 1)) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "none") +
  xlab("") +
  ylab("") +
  scale_x_reverse()

7

Code

df.plotComp7 %>%
  mutate(svar = as.numeric(svarskategori)) %>% 
  mutate(item = factor(item)) %>% 
  #mutate(item = fct_rev(item)) %>% 
  ggplot(
    aes(
      x = svar, fill = group, color = group
    )) +
  geom_density(alpha = 0.6, stat = "count") +
  facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             ) +
  scale_x_continuous(limits = svarskategorier, 
                     breaks = svarskategorier,
                     labels = svarskategorier
    ) +
  scale_y_continuous(position="right", 
                     minor_breaks = NULL,
                     label = scales::label_comma(accuracy = 1)) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "none") +
  xlab("") +
  ylab("") +
  scale_x_reverse()

40

Code

df.plotComp40 %>%
  mutate(svar = as.numeric(svarskategori)) %>% 
  mutate(item = factor(item)) %>% 
  #mutate(item = fct_rev(item)) %>% 
  ggplot(
    aes(
      x = svar, fill = group, color = group
    )) +
  geom_density(alpha = 0.6, stat = "count") +
  facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             ) +
  scale_x_continuous(limits = svarskategorier, 
                     breaks = svarskategorier,
                     labels = svarskategorier
    ) +
  scale_y_continuous(position="right", 
                     minor_breaks = NULL,
                     label = scales::label_comma(accuracy = 1)) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "none") +
  xlab("") +
  ylab("") +
  scale_x_reverse()

14.6.3 columns + density

20

Code

# same but without y axis text
colComp20 <- df.plotComp20 %>%
  group_by(group) %>% 
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = n, fill = group
    )
  ) +
  geom_col(position = "dodge") +
  scale_color_manual("",
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
    scale_y_continuous(position = "right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             #scales = "free"
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "top",
                axis.text.y = element_blank()) +  
  xlab("") +
  ylab("") +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))
#colComp20

Code

## densityplot with no labels
densityPlot20 <- df.plotComp20 %>%
  mutate(svar = as.numeric(svarskategori)) %>% 
  mutate(item = factor(item)) %>% 
  #mutate(item = fct_rev(item)) %>% 
  ggplot(
    aes(
      x = svar, fill = group, color = group
    )) +
  geom_density(alpha = 0.6, stat = "count") +
  facet_wrap(~ item,
             dir = "v",
             ncol = 1) +
  scale_x_continuous(limits = c(1,6), 
                     breaks = c(1:6)
                     ) +
  scale_y_continuous(minor_breaks = NULL) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal(base_family = "noto") +
  theme_prevent(stripsize = 9) +
  xlab("") +
  ylab("") +
  theme(legend.position = "none") +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank(),
    strip.text = element_blank()
  ) +
  scale_x_reverse()

#densityPlot

Code

p20cols <- colComp20 + 
  theme(plot.margin = unit(c(0,0,0,0), 'cm'))
p20dens <- densityPlot20 + 
  theme(plot.margin = unit(c(0,0,0,0), 'cm'))
# unit(c(top, right, bottom, left), units)
p20cols + p20dens + plot_layout(widths = c(2.3, 1))

7

Code

# same but without y axis text
colComp7 <- df.plotComp7 %>%
  group_by(group) %>% 
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = n, fill = group
    )
  ) +
  geom_col(position = "dodge") +
  scale_color_manual("",
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
    scale_y_continuous(position = "right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             #scales = "free"
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "top",
                axis.text.y = element_blank()) +  
  xlab("") +
  ylab("") +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))
#colComp20

Code

## densityplot with no labels
densityPlot7 <- df.plotComp7 %>%
  mutate(svar = as.numeric(svarskategori)) %>% 
  mutate(item = factor(item)) %>% 
  #mutate(item = fct_rev(item)) %>% 
  ggplot(
    aes(
      x = svar, fill = group, color = group
    )) +
  geom_density(alpha = 0.6, stat = "count") +
  facet_wrap(~ item,
             dir = "v",
             ncol = 1) +
  scale_x_continuous(limits = c(1,6), 
                     breaks = c(1:6)
                     ) +
  scale_y_continuous(minor_breaks = NULL) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal(base_family = "noto") +
  theme_prevent(stripsize = 9) +
  xlab("") +
  ylab("") +
  theme(legend.position = "none") +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank(),
    strip.text = element_blank()
  ) +
  scale_x_reverse()

#densityPlot

Code

p7cols <- colComp7 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
p7dens <- densityPlot7 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
# unit(c(top, right, bottom, left), units)
p7cols + p7dens + plot_layout(widths = c(2.3, 1))

40

Code

# same but without y axis text
colComp40 <- df.plotComp40 %>%
  group_by(group) %>% 
  dplyr::count(item, svarskategori, .drop = F) %>%
  mutate(Nf = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = n, fill = group
    )
  ) +
  geom_col(position = "dodge") +
  scale_color_manual("",
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
    scale_y_continuous(position = "right",
                     label = scales::label_comma(accuracy = 1),
                     minor_breaks = NULL) +
    facet_wrap(~ item,
             dir = "v",
             ncol = 1,
             strip.position = "left",
             labeller = labeller(item = label_wrap_gen(22))
             #scales = "free"
             ) +
  theme_minimal() +
  theme_prevent(stripsize = 10, 
                strip.text.y.left = element_text(angle = 0),
                legend.position = "top",
                axis.text.y = element_blank()) +  
  xlab("") +
  ylab("") +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))
#colComp20

Code

## densityplot with no labels
densityPlot40 <- df.plotComp40 %>%
  mutate(svar = as.numeric(svarskategori)) %>% 
  mutate(item = factor(item)) %>% 
  #mutate(item = fct_rev(item)) %>% 
  ggplot(
    aes(
      x = svar, fill = group, color = group
    )) +
  geom_density(alpha = 0.6, stat = "count") +
  facet_wrap(~ item,
             dir = "v",
             ncol = 1) +
  scale_x_continuous(limits = c(1,6), 
                     breaks = c(1:6)
                     ) +
  scale_y_continuous(minor_breaks = NULL) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal(base_family = "noto") +
  theme_prevent(stripsize = 9) +
  xlab("") +
  ylab("") +
  theme(legend.position = "none") +
  theme(
    axis.text.x = element_blank(), # remove text from both axes
    axis.text.y = element_blank(),
    axis.title = element_blank(),
    strip.text = element_blank()
  ) +
  scale_x_reverse()

#densityPlot

Code

p40cols <- colComp40 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
p40dens <- densityPlot40 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
# unit(c(top, right, bottom, left), units)
p40cols + p40dens + plot_layout(widths = c(2.3, 1))

14.6.4 stat_dots

20

Code

statDots20 <- df.plotComp20 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      color = group, fill = group
    )
  ) +
  stat_dots(data = . %>% filter(group == "Mättillfälle 1"),
            position = position_nudge(y = 0.04),
            #scale = 0.5
            #dotsize = 1.6
            ) +
    stat_dots(data = . %>% filter(group == "Mättillfälle 2"),
            position = position_nudge(y = -0.04),
            #scale = 0.9
            #dotsize = 2
            ) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8)) +
  xlab("") +
  ylab("") +
  theme(legend.position = "top")

statDots20

7

Code

statDots7 <- df.plotComp7 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      color = group, fill = group
    )
  ) +
  stat_dots(data = . %>% filter(group == "Mättillfälle 1"),
            position = position_nudge(y = 0.04),
            #scale = 0.5
            #dotsize = 1.6
            ) +
    stat_dots(data = . %>% filter(group == "Mättillfälle 2"),
            position = position_nudge(y = -0.04),
            #scale = 0.9
            #dotsize = 2
            ) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8)) +
  xlab("") +
  ylab("") +
  theme(legend.position = "top")

statDots7

40

Code

statDots40 <- df.plotComp40 %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  ggplot(
    aes(
      x = svarskategori, y = item,
      color = group, fill = group
    )
  ) +
  stat_dots(data = . %>% filter(group == "Mättillfälle 1"),
            position = position_nudge(y = 0.04),
            #scale = 0.5
            #dotsize = 1.6
            ) +
    stat_dots(data = . %>% filter(group == "Mättillfälle 2"),
            position = position_nudge(y = -0.04),
            #scale = 0.9
            #dotsize = 2
            ) +
  scale_color_manual('Grupp',
    values = c(prevent_light_blue,prevent_yellow),
    aesthetics = c("fill", "color")
  ) +
  theme_minimal(base_family = "noto") +
  theme_prevent() +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8)) +
  xlab("") +
  ylab("") +
  theme(legend.position = "top")

statDots40

14.6.5 stat_dots + density

20

Code

p1 <- statDots20 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
p2 <- densityPlot20 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
# unit(c(top, right, bottom, left), units)
p1 + p2 + plot_layout(widths = c(2.5, 1))

7

Code

p1 <- statDots7 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
p2 <- densityPlot7 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
# unit(c(top, right, bottom, left), units)
p1 + p2 + plot_layout(widths = c(2.5, 1))

40

Code

p1 <- statDots40 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
p2 <- densityPlot40 + theme(plot.margin = unit(c(0,0,0,0), 'cm'))
# unit(c(top, right, bottom, left), units)
p1 + p2 + plot_layout(widths = c(2.5, 1))

14.6.6 ridges

20

Code

#densityRidgePlot <- 
  
df.plotComp20 %>%
  mutate(svar = as.numeric(svarskategori)) %>%
  mutate(item = fct_rev(item)) %>%
  ggplot(
    aes(
      x = svar, y = item,
      fill = group
    )) +
  geom_density_ridges(alpha = 0.6,
                      scale = 1.5) +
  scale_x_continuous('',
                     labels = svarskategorier,
                     limits = c(1,6),
                     breaks = c(1:6)) +
  scale_fill_cyclical('Grupp',
    values = c(prevent_light_blue,prevent_yellow)
  ) +
  theme_minimal() +
  theme_prevent() +
  xlab("") +
  ylab("") +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30))

Code

  #scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))


#densityRidgePlot

7

Code

#densityRidgePlot <- 
  
df.plotComp7 %>%
  mutate(svar = as.numeric(svarskategori)) %>%
  mutate(item = fct_rev(item)) %>%
  ggplot(
    aes(
      x = svar, y = item,
      fill = group
    )) +
  geom_density_ridges(alpha = 0.6,
                      scale = 1.5) +
  scale_x_continuous('',
                     labels = svarskategorier,
                     limits = c(1,6),
                     breaks = c(1:6)) +
  scale_fill_cyclical('Grupp',
    values = c(prevent_light_blue,prevent_yellow)
  ) +
  theme_minimal() +
  theme_prevent() +
  xlab("") +
  ylab("") +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30))

40

Code

#densityRidgePlot <- 
  
df.plotComp40 %>%
  mutate(svar = as.numeric(svarskategori)) %>%
  mutate(item = fct_rev(item)) %>%
  ggplot(
    aes(
      x = svar, y = item,
      fill = group
    )) +
  geom_density_ridges(alpha = 0.6,
                      scale = 1.5) +
  scale_x_continuous('',
                     labels = svarskategorier,
                     limits = c(1,6),
                     breaks = c(1:6)) +
  scale_fill_cyclical('Grupp',
    values = c(prevent_light_blue,prevent_yellow)
  ) +
  theme_minimal() +
  theme_prevent() +
  xlab("") +
  ylab("") +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30))

14.6.7 geom_point

20

Code

df.pointComp <- df.plotComp20 %>%
  dplyr::count(group, item, svarskategori) %>%
  mutate(item = fct_rev(item)) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori))

ggplot() +
  # layer for previous measurement
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    alpha = 0.3,
    shape = 16,
    position = position_nudge(x = 0.3, y = 0)
  ) +
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    # size = 3,
    shape = 16
  ) +
  # create outline for each point
  #     geom_point(data = df.pointComp %>% filter(group == "Mättillfälle 1"),
  #   aes(x = svarskategori, y = item, size = n * 1.5),
  #   color = "lightgrey",
  #   # size = 3,
  #   shape = 1
  # ) +
  # text layer for showing number of responses in each category
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, label = n),
    color = "white"
  ) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none" # remove legend for size aesthetic
  ) +
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, label = n),
    color = "grey",
    position = position_nudge(x = 0.3, y = 0)
  ) +
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "top") +
  scale_color_manual('',
    values = c(prevent_light_blue,prevent_yellow)
                     ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

20 grön skala

Code

df.pointComp <- df.plotComp20 %>%
  dplyr::count(group, item, svarskategori) %>%
  mutate(item = fct_rev(item)) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori))

ggplot() +
  # layer for previous measurement
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    alpha = 0.3,
    shape = 16,
    position = position_nudge(x = 0.3, y = 0)
  ) +
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  # create outline for each point
  #     geom_point(data = df.pointComp %>% filter(group == "Mättillfälle 1"),
  #   aes(x = svarskategori, y = item, size = n * 1.5),
  #   color = "lightgrey",
  #   # size = 3,
  #   shape = 1
  # ) +
  # text layer for showing number of responses in each category
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, label = n),
    color = "white"
  ) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none" # remove legend for size aesthetic
  ) +
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, label = n),
    color = "grey",
    position = position_nudge(x = 0.3, y = 0)
  ) +
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "top") +
  scale_color_viridis_d("",
    begin = 0.2,
    end = 0.8,
    guide = "none"
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

Code

df.pointComp <- df.plotComp20 %>%
  dplyr::count(group, item, svarskategori) %>%
  mutate(item = fct_rev(item)) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  mutate(svarNum = as.numeric(svarskategori)) %>% 
  mutate(svarSumma = svarNum * n)

df.groupdiff <- df.pointComp %>% 
  group_by(group, item) %>% 
  summarise(groupSum = sum(svarSumma)) %>% 
  pivot_wider(names_from = "group", values_from = "groupSum") %>% 
  janitor::clean_names() %>% 
  mutate(groupDiff = mattillfalle_2 - mattillfalle_1) # positive value = positive development
  
  # ggplot() +
  # geom_col(aes(x = item, y = groupSum, group = group, fill = group),
  #          position = "dodge")
library(metR)
plotPrePost <- ggplot() +
  # layer for previous measurement
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    alpha = 0.3,
    shape = 16,
    position = position_nudge(x = 0.3, y = 0)
  ) +
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    # size = 3,
    shape = 16
  ) +
  # create outline for each point
  #     geom_point(data = df.pointComp %>% filter(group == "Mättillfälle 1"),
  #   aes(x = svarskategori, y = item, size = n * 1.5),
  #   color = "lightgrey",
  #   # size = 3,
  #   shape = 1
  # ) +
  # text layer for showing number of responses in each category
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, label = n),
    color = "white"
  ) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none" # remove legend for size aesthetic
  ) +
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, label = n),
    color = "grey",
    position = position_nudge(x = 0.3, y = 0)
  ) +
  # trend symbol
  geom_arrow(
    data = df.groupdiff,
    aes(x = 0.4, y = item, angle = 0 + groupDiff, mag = 5),
    show.legend = F
  ) +
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "top") +
  scale_color_manual('',
    values = c(prevent_light_blue,prevent_yellow)
                     ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 22)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

#plotPrePost

20 med density

Code

plotPrePost + densityPlot20 + plot_layout(widths = c(2.5, 1))

20 grön med trend-pil

Code

df.pointComp <- df.plotComp20 %>%
  dplyr::count(group, item, svarskategori) %>%
  mutate(item = fct_rev(item)) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori)) %>% 
  mutate(svarNum = as.numeric(svarskategori)) %>% 
  mutate(svarSumma = svarNum * n)

df.groupdiff <- df.pointComp %>% 
  group_by(group, item) %>% 
  summarise(groupSum = sum(svarSumma)) %>% 
  pivot_wider(names_from = "group", values_from = "groupSum") %>% 
  janitor::clean_names() %>% 
  mutate(groupDiff = mattillfalle_2 - mattillfalle_1) # positive value = positive development
  
  # ggplot() +
  # geom_col(aes(x = item, y = groupSum, group = group, fill = group),
  #          position = "dodge")
library(metR)
ggplot() +
  # layer for previous measurement
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    alpha = 0.3,
    shape = 16,
    position = position_nudge(x = 0.3, y = 0)
  ) +
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = svarskategori),
    # size = 3,
    shape = 16
  ) +
  # create outline for each point
  #     geom_point(data = df.pointComp %>% filter(group == "Mättillfälle 1"),
  #   aes(x = svarskategori, y = item, size = n * 1.5),
  #   color = "lightgrey",
  #   # size = 3,
  #   shape = 1
  # ) +
  # text layer for showing number of responses in each category
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, label = n),
    color = "white"
  ) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none" # remove legend for size aesthetic
  ) +
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, label = n),
    color = "grey",
    position = position_nudge(x = 0.3, y = 0)
  ) +
  # trend symbol
  geom_arrow(
    data = df.groupdiff,
    aes(x = 0.4, y = item, angle = 0 + groupDiff, mag = 5),
    show.legend = F
  ) +
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "top") +
  scale_color_viridis_d("",
    begin = 0.2,
    end = 0.8,
    guide = "none"
  ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

7

Code

df.pointComp <- df.plotComp7 %>%
  dplyr::count(group, item, svarskategori) %>%
  mutate(item = fct_rev(item)) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori))

ggplot() +
  # layer for previous measurement
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    alpha = 0.3,
    shape = 16,
    position = position_nudge(x = 0.3, y = 0)
  ) +
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    # size = 3,
    shape = 16
  ) +
  # create outline for each point
  #     geom_point(data = df.pointComp %>% filter(group == "Mättillfälle 1"),
  #   aes(x = svarskategori, y = item, size = n * 1.5),
  #   color = "lightgrey",
  #   # size = 3,
  #   shape = 1
  # ) +
  # text layer for showing number of responses in each category
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, label = n),
    color = "white"
  ) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none" # remove legend for size aesthetic
  ) +
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, label = n),
    color = "grey",
    position = position_nudge(x = 0.3, y = 0)
  ) +
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "top") +
  scale_color_manual('',
    values = c(prevent_light_blue,prevent_yellow)
                     ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))

40

Code

df.pointComp <- df.plotComp40 %>%
  dplyr::count(group, item, svarskategori) %>%
  mutate(item = fct_rev(item)) %>%
  mutate(nFactor = factor(n)) %>%
  mutate(svarskategori = fct_rev(svarskategori))

ggplot() +
  # layer for previous measurement
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    alpha = 0.3,
    shape = 16,
    position = position_nudge(x = 0.3, y = 0)
  ) +
  geom_point(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, size = n * 1.5, color = group),
    # size = 3,
    shape = 16
  ) +
  # create outline for each point
  #     geom_point(data = df.pointComp %>% filter(group == "Mättillfälle 1"),
  #   aes(x = svarskategori, y = item, size = n * 1.5),
  #   color = "lightgrey",
  #   # size = 3,
  #   shape = 1
  # ) +
  # text layer for showing number of responses in each category
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 2"),
    aes(x = svarskategori, y = item, label = n),
    color = "white"
  ) +
  scale_size_continuous(
    range = c(7, 20), # set minimum and maximum point size
    guide = "none" # remove legend for size aesthetic
  ) +
  geom_text(
    data = df.pointComp %>% filter(group == "Mättillfälle 1"),
    aes(x = svarskategori, y = item, label = n),
    color = "grey",
    position = position_nudge(x = 0.3, y = 0)
  ) +
  ### theming, colors, fonts, etc below
  theme_minimal() +
  theme_prevent(legend.position = "top") +
  scale_color_manual('',
    values = c(prevent_light_blue,prevent_yellow)
                     ) +
  labs(
    title = "Indexfrågor",
    subtitle = "Fördelning av svar",
    y = "",
    x = ""
  ) +
  scale_y_discrete(labels = ~ stringr::str_wrap(.x, width = 30)) +
  scale_x_discrete(labels = ~ stringr::str_wrap(.x, width = 8))