Data Visualization Guidelines

R is an amazing tool for data analysis and visualization. The code on this page shows the a variety of plots using the R ggplot2 package to display statistical outputs. All the data shown on the page is fictitious and only meant for illustrative purposes.

Use of Fonts

• Century Gothic Bold, font size 12 should be used for the labels and values on the graph. - Download here

General Rules

• All axes should be labelled and where possible, the values for data points should be indicated on the graph.
• Prevent use of grid lines whenever possible to not clutter plots
• Use line breaks in long labels
• Be conscious about the way axis are ordered. For example, order regions either from the lowest to the highest value or using the serpentine order.
• label axis appropriately

Use of Colours

Please use the following colour schemes for different types of disaggregation.

Sex

Male

Hex: #206095
rgb(32, 96, 149)
.

Female

Hex: #F66068
rgb(246, 96,104)
.

Locality Type

National

#27A0CC
.

Urban

#871A5B
.

Rural

#22D0B6
.

Economic Sectors

Industry

#14607A
.

Agriculture

#07BB9E
.

Service

#F98B00
.

Neutral

Neutral is used on general variables. It is advise that the maximum number of variables per plot should not be more than 5.

#002060

#0070C0

#00B0F0

#8EA9DB

#9BC2E6

#FFFFCC

Positive ~ Negative

#6AA84F

#93C47D

#FFFFCC

#F4CCCC

#E06666

Population & Density

#FFFFCC

#C7E9B4

#7FCDBB

#41B6C4

#2C7FB8

Incidence

#FECCCC

#FF9999

#FF6666

#FF3333

#CC0000

#990000

Food / Non-food

Food

#3ECDB9

Non-food

#04BCFC

16 Regions and generic colours

These colour can be used for the 16 different regions or to assign more generic colours to a visualization that uses categories that are not defined in this visualization guide.

Ahafo

#4B644B

Ashanti

#7D96AF

Bono

#EBA07E

Bono East

#09979B

Central

#EA879C

Eastern

#E5B9AD

Greater Accra

#CC9B7A

Northern

#FDD835

North East

#0070C0

Oti

#AE2B47

Savannah

#F94240

Upper East

#903000

Upper West

#0F3B6C

Volta

#59A77F

Western

#FDAE6B

Western North

#B173A0

Missing value

If you want to assign a colour to missing values, you can use this grey tone.

Missing

#ADABAC

Using GSS theme

In the future, GSS will release it is own package for data visualizations. For now it suffices to run this code to load the all the R code to set the right defaults (colours, fonts, etc.) to create plots in the GGS theme.

Show the code

# load packages
library(ggplot2)
library(tidyverse)
library(ggtext)
library(gghighlight)
library(sysfonts)
library(showtextdb)
library(showtext)
library(glue)
library(scales)
library(kableExtra)
library(patchwork)
library(forcats)
library(tidyverse)
library(ggbump)
library(reshape2)
library(plotly)


# load the font
font_add("century gothic bold", "Font/Century Gothic.ttf")

# make sure ggplot recognizes the font 
# and set the font to high-res
showtext_auto()
showtext::showtext_opts(dpi = 300)

# set default colour for plots with multiple categories
options(ggplot2.discrete.colour = c("#210D69", "#DB2E76", "#586889", "#227C42"))
options(ggplot2.discrete.fill   = c("#210D69", "#DB2E76", "#586889", "#227C42"))

# set default colour for plots with a single category
update_geom_defaults("bar",   list(fill = "#27A0CC"))
update_geom_defaults("col",   list(fill = "#27A0CC"))

# update the font to show in geom_text()
update_geom_defaults("text",   list(family = "century gothic bold", size = 4.5 ))
GSS_font <- "century gothic bold"
# create a GGS theme based on the theme_gray()
gssthemes<-function(){
  theme_gray() %+replace%
    theme(
      text=element_text(family = "century gothic bold",
                        colour="black",
                        size=10),
      plot.margin = margin(0.5,0.3, 0.3, 0.3, "cm"),
      # plot.title =element_textbox_simple(family="century gothic bold", size=16,
      #                                    lineheight=1,
      #                                    margin=margin(b=10)),
      # plot.title.position="plot",
      plot.caption=element_markdown(hjust=0, color="gray",
                                    lineheight=1.5,
                                    margin =margin(t=10)),
      plot.caption.position="plot",
      axis.title.y=element_text(color="black", angle=90, size = 10),
      axis.title.x=element_text(color="black",size = 10),
      axis.text.x=element_text(color="black", size = 10, vjust = 0, margin = margin(t = 5, r = 5, b = 0, l = 0, unit = "pt")),
      axis.text.y=element_text(color="black", size = 10, hjust = 1, margin = margin(t = 5, r = 5, b = 0, l = 0, unit = "pt")),
      legend.text=element_text(color="black",  size = 10),
      panel.grid.major.y=element_line(color="gray", size=0.25),
      panel.grid.major.x=element_blank(),
      panel.grid.minor=element_blank(),
      panel.background=element_rect(fill="white", color=NA),
      plot.background=element_rect(fill="white", color=NA),
      legend.background=element_rect(fill="white", color=NA),
     strip.background =element_rect(fill="white",  color=NA),
     legend.key = element_rect(fill = "white", color = NA), 
     strip.text = element_text( size = 20,  margin = margin(t = 5, r = 0, b = 10, l = 0, unit = "pt"))
    )
}

# this will make the labels of the bar chart a bit nicer, by ending above the highest data point
nicelimits <- function(x) {
  range(scales::extended_breaks(only.loose = TRUE)(x))
}


# Define color palette

statscolours_color_scheme  <- c("#382873", "#0168C8", "#00B050")

#locality
national_color <- "#27A0CC"
urban_color <- "#871A5B"
rural_color <- "#22D0B6"
urbanrural_color_scheme  <- c("#27A0CC", "#871A5B", "#22D0B6")

#Sex
male_color <- "#206095"
female_color <- "#F66068"
malefemale_color_scheme <- c("#27A0CC", "#206095", "#F66068")

#positive/negative
negative_color <- "#cc3333"
positive_color <- "#33cccc"

#pallete
neutral_color_scheme  <- c("#002060", "#0070C0", "#00B0F0", "#8EA9DB", "#9BC2E6", "#FFFFCC")
posneg_color_scheme  <- c("#38761D","#6AA84F","#93C47D","#F4CCCC","#E06666","#990000")
posneutralneg_color_scheme  <- c("#38761D","#6AA84F","#FFFFCC","#E06666","#990000")
population_color_scheme  <- c("#FFFFCC","#C7E9B4","#7FCDBB","#41B6C4","#2C7FB8")
incidence_color_scheme  <- c("#FECCCC","#FF9999","#FF6666","#FF3333","#CC0000","#990000")


#Economic sector colour
industry_color<-"#14607A"
agric_color<-"#07BB9E"
services_color<-"#F98B00"
economic_sectors<- c("#14607A","#07BB9E","#F98B00")

#food
food_colour<- "#3ECDB9"
nonfood_colour<-"#04BCFC"

# Region
Ahafo_color_code <- "#4B644B"
Ashanti_color_code <- "#7D96AF"
Bono_color_code <- "#EBA07E"
Bono_East_color_code <- "#09979B"
Central_color_code <- "#EA879C"
Eastern_color_code <- "#E5B9AD"
Greater_Accra_color_code <- "#CC9B7A"
Northern_color_code <- "#FDD835"
North_East_code <- "#0070C0"
Oti_color_code <- "#AE2B47"
Savannah_color_code <- "#F94240"
Upper_East_color_code <- "#903000"
Upper_West_color_code <- "#0F3B6C"
Volta_color_code <- "#59A77F"
Western_color_code <- "#FDAE6B"
Western_North_color_code <- "#B173A0"

Bar charts

A bar chart is an effective way to visually represent data that is categorical or discrete in nature for example different regions in Ghana). It is particularly useful when comparing values across different categories or groups. Bar charts are ideal for showing the distribution or frequency of data, as well as identifying trends or patterns over time. They can be used to display numerical data such as quantities, percentages, and proportions. Overall, a bar chart is appropriate when you want to easily compare different categories or groups and understand the relative differences between them.

example data

Show the code

n_chopbars_df  <- tribble(~region_number, ~region, ~number_of_chop_bars, ~urban, ~rural,
                            1,"Western", 567, 300, 267,
                            2, "Central", 907, 600, 307,
                            3, "Greater Accra", 6990, 4792, 2198, 
                            4, "Volta", 278, 78, 200, 
                            5, "Eastern", 787, 287, 500,
                            6, "Ashanti",5621, 621, 5000,
                            7, "Western North", 345, 145, 200,
                            8,  "Ahafo", 678, 339, 339,
                            9, "Bono",1268, 600, 668, 
                            10,"Bono East", 1268,1048, 220,
                            11,"Oti",773, 500, 273, 
                            12,"Northern", 112, 12, 100,
                            13,"Savannah", 1289,503, 786, 
                            14,"North East",378, 142, 236,
                            15,"Upper East",1279,326, 953, 
                            16, "Upper West", 3782, 3456, 326) 

n_chopbars_df %>%
    kable() %>%
    kable_styling() %>% 
  scroll_box(height = "400px")

region_number	region	number_of_chop_bars	urban	rural
1	Western	567	300	267
2	Central	907	600	307
3	Greater Accra	6990	4792	2198
4	Volta	278	78	200
5	Eastern	787	287	500
6	Ashanti	5621	621	5000
7	Western North	345	145	200
8	Ahafo	678	339	339
9	Bono	1268	600	668
10	Bono East	1268	1048	220
11	Oti	773	500	273
12	Northern	112	12	100
13	Savannah	1289	503	786
14	North East	378	142	236
15	Upper East	1279	326	953
16	Upper West	3782	3456	326

One Color

Show the code

# Pipe operator to pass data frame to the next line
n_chopbars_df %>% 
  # Create a ggplot object and set the aesthetic mappings
  ggplot(mapping = aes(x = region, y = number_of_chop_bars)) +
  # Add a column chart with bars of equal width
  geom_col(width = 0.8) +
  # Apply a custom theme to the plot
  gssthemes() +
# The expand argument controls whether the range of the y-axis is expanded to include a small margin around the data. The limits argument sets the upper and lower limits of the y-axis to nicelimits, which is a function that makes sure the limits are always above the largest data point. he breaks argument sets the tick marks on the y-axis to use extended_breaks from the scales package, which generates a sequence of evenly spaced values with loose spacing.
  scale_y_continuous( expand = c( 0, 1 ),
                      limits = nicelimits,
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    # Set the scale for the x-axis with tick labels rotated by 90 degrees
  scale_x_discrete(guide = guide_axis(angle = 90)) +
     # Add axis labels to the plot
    labs(x = NULL,
             y = "number of chop bars")+
    # Set the coordinate system for the plot, allowing data points to be partially displayed outside of the plot area.
  coord_cartesian(clip = "off")

One Color (Rotated) with labels

This example introduces coord_flip() instead of coord_cartesian() and bring back some theme elements to draw vertical instead of horizontal grid lines.

Show the code

# Pipe operator to pass data frame to the next line
n_chopbars_df %>% 
  # Create a ggplot object and set the aesthetic mappings
  ggplot(mapping = aes(x = region, y = number_of_chop_bars)) +
  # Add a column chart with bars of equal width
  geom_col(width = 0.8) +
    # add text to the end of the plot
     geom_text(mapping = aes(label = number_of_chop_bars), hjust = -0.2) +
  # Apply a custom theme to the plot
  gssthemes() +
    # The expand argument controls whether the range of the y-axis is expanded to include a small margin around the data. 
    # The limits argument sets the upper and lower limits of the y-axis to nicelimits, which is a function that makes sure 
    # the limits are always above the largest data point. he breaks argument sets the tick marks on the y-axis to use
    # extended_breaks from the scales package, which generates a sequence of evenly spaced values with loose spacing.
  scale_y_continuous( expand = c( 0, 1 ),
                      limits = nicelimits,
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    # Add axis labels to the plot
    labs(x = NULL,
             y = "number of chop bars")+
    # Set the coordinate system for the plot, allowing data points to be partially displayed outside of the plot area.
  coord_flip(clip = "off") +
    # change the theme a bit so that 1) the axis lines are vertical and 2) the labels are right alligend
    theme(panel.grid.major.x=element_line(color = "gray", size=0.25),
                panel.grid.major.y=element_blank(),
                axis.text.x = element_text(vjust = 0.5))

One Color (Rotated) with labels and ordered

You can use the reorder() function to order the axis labels.

Show the code

# Pipe operator to pass data frame to the next line
n_chopbars_df %>% 
  # Create a ggplot object and set the aesthetic mappings
    # use the reorder function to reorder the the bars
  ggplot(mapping = aes(x = reorder(region,number_of_chop_bars), y = number_of_chop_bars)) +
  # Add a column chart with bars of equal width
  geom_col(width = 0.8) +
    # add text to the end of the plot, and comma at all thousands
     geom_text(mapping = aes(label = scales::comma(number_of_chop_bars)), hjust = -0.2,) +
  # Apply a custom theme to the plot
  gssthemes() +
    # The expand argument controls whether the range of the y-axis is expanded to include a small margin around the data. 
    # The limits argument sets the upper and lower limits of the y-axis to nicelimits, which is a function that makes sure 
    # the limits are always above the largest data point. he breaks argument sets the tick marks on the y-axis to use
    # extended_breaks from the scales package, which generates a sequence of evenly spaced values with loose spacing.
  scale_y_continuous( expand = c( 0, 1 ),
                      limits = nicelimits,
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    # Add axis labels to the plot
    labs(x = NULL,
             y = "number of chop bars")+
    # Set the coordinate system for the plot, allowing data points to be partially displayed outside of the plot area.
  coord_flip(clip = "off") +
    # change the theme a bit so that 1) the axis lines are vertical and 2) the labels are right alligend
    theme(panel.grid.major.x=element_line(color = "gray", size=0.25),
                panel.grid.major.y=element_blank(),
                axis.text.x = element_text(vjust = 0.5))

Show different categories using color palettes (stacked)

If you want to show different categories, you can use the fill = categorical variable argument inside the aesthetics mapping. To do this, the data needs to be in a long format

example data (long format)

Show the code

n_chopbars_df_longformat <- n_chopbars_df %>% 
    select(-number_of_chop_bars ) %>% 
    pivot_longer(-c(region_number, region), names_to = "locality", values_to = "number_of_chop_bars")

n_chopbars_df_longformat %>%
    kable() %>%
    kable_styling() %>% 
  scroll_box(height = "400px")

region_number	region	locality	number_of_chop_bars
1	Western	urban	300
1	Western	rural	267
2	Central	urban	600
2	Central	rural	307
3	Greater Accra	urban	4792
3	Greater Accra	rural	2198
4	Volta	urban	78
4	Volta	rural	200
5	Eastern	urban	287
5	Eastern	rural	500
6	Ashanti	urban	621
6	Ashanti	rural	5000
7	Western North	urban	145
7	Western North	rural	200
8	Ahafo	urban	339
8	Ahafo	rural	339
9	Bono	urban	600
9	Bono	rural	668
10	Bono East	urban	1048
10	Bono East	rural	220
11	Oti	urban	500
11	Oti	rural	273
12	Northern	urban	12
12	Northern	rural	100
13	Savannah	urban	503
13	Savannah	rural	786
14	North East	urban	142
14	North East	rural	236
15	Upper East	urban	326
15	Upper East	rural	953
16	Upper West	urban	3456
16	Upper West	rural	326

Show the code

n_chopbars_df_longformat %>% 
  ggplot(mapping = aes(x = reorder(region,number_of_chop_bars), fill = locality ,  y = number_of_chop_bars)) +
  geom_col(width = 0.8) +
  gssthemes() +
  scale_y_continuous( expand = c( 0, 1 ),
                      limits = nicelimits,
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    labs(x = NULL,
             y = "number of chop bars")+
  coord_flip(clip = "off") +
    scale_fill_manual(values = c(urban = urban_color,
                                                             rural = rural_color))+
    theme(panel.grid.major.x=element_line(color = "gray", size=0.25),
                panel.grid.major.y=element_blank(),
                axis.text.x = element_text(vjust = 0.5))

Show different categories using color palettes (dodged)

Instead of having the categories stacked, you can put them next to each other (stacked)

Show the code

n_chopbars_df_longformat %>% 
  ggplot(mapping = aes(x = reorder(region,number_of_chop_bars), fill = locality ,  y = number_of_chop_bars)) +
  geom_col(width = 0.8, position = "dodge") +
    geom_text(mapping = aes(label = scales::comma(number_of_chop_bars)), hjust = -0.2, 
                        position = position_dodge(width = 0.8)) +
  gssthemes() +
  scale_y_continuous( expand = c( 0, 0 ),
                      limits = nicelimits,
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    labs(x = NULL,
             y = "number of chop bars")+
  coord_flip(clip = "off") +
    scale_fill_manual(values = c(urban = urban_color,
                                                             rural = rural_color))+
    theme(panel.grid.major.x=element_line(color = "gray", size=0.25),
                panel.grid.major.y=element_blank(),
                axis.text.x = element_text(vjust = 0.5))

Bar chart with highlighted category

Sometimes you might want to highlight a single category in a bar chart (for example the National/Ghana average). To highlight also the axis label you need to use the showtext package. It is recommended to put highlight is a similar hue than the other colours. This can be done, using the muted() function from the scales package

Show the code

# this code add the Ghana average
n_chopbars_df_with_national <- n_chopbars_df %>% 
    add_row( region_number = 0,
                         region  = "Ghana",
                         number_of_chop_bars = mean(n_chopbars_df$number_of_chop_bars),
                         urban = mean(n_chopbars_df$urban ),
                         rural= mean(n_chopbars_df$rural) )

Show the code

highlight = function(x, pat, color="black", family="") {
    ifelse(grepl(pat, x), glue("<b style='font-family:{family}; color:{color}'>{x}</b>"), x)
}

# Pipe operator to pass data frame to the next line
n_chopbars_df_with_national %>% 
    mutate(highlight = region == "Ghana") %>% 
  # Create a ggplot object and set the aesthetic mappings
    # use the reorder function to reorder the the bars
  ggplot(mapping = aes(x = reorder(region,number_of_chop_bars), 
                                         y = number_of_chop_bars,
                                          fill = highlight)) +
  geom_col(width = 0.8) +
    geom_text(mapping = aes(label = scales::comma(number_of_chop_bars)), hjust = -0.2) +
  gssthemes() +
  scale_y_continuous( expand = c( 0, 0 ),
                      limits = nicelimits,
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    scale_x_discrete(labels= function(x) highlight(x, "Ghana", "black")) +
    scale_fill_manual(values = c(`FALSE` = national_color, 
                                                             `TRUE` = scales::muted(national_color)),
                                        guide = "none") +
    labs(x = NULL,
             y = "number of chop bars" )+
  coord_flip(clip = "off") +
    theme(panel.grid.major.x = element_line(color = "gray", size=0.25),
                panel.grid.major.y = element_blank(),
                axis.text.x = element_text(vjust = 0.5),
                axis.text.y = element_markdown()) # to make this work you need to make the y axis a markdown format

Bar chart with long labels

To allow for sufficient plotting space, it is a good idea to add line break to long value labels on the axes. You can use the str_wrap() function to add line breaks

example data (long labels)

Show the code

n_chopbars_df_long_labels  <- tribble(~question, ~count, 
                                                                            "number of households eating at chopbars every day", 678,
                                                                            "number of male headed households eating at chopbars once a week", 1267,
                                                                            "number of female headed households eating at chop bars once a week", 1876)
n_chopbars_df_long_labels %>%
    kable() %>%
    kable_styling()

question	count
number of households eating at chopbars every day	678
number of male headed households eating at chopbars once a week	1267
number of female headed households eating at chop bars once a week	1876

Show the code

n_chopbars_df_long_labels %>% 
  ggplot(mapping = aes(x = str_wrap(question, 20), y = count)) +
  geom_col(width = 0.8) +
    geom_text(mapping = aes(label = scales::comma(count)), hjust = -0.2) +
  gssthemes() +
  scale_y_continuous( expand = c( 0, 1 ),
                      limits = nicelimits,
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    labs(x = NULL,
             y = "number of chop bars")+
  coord_flip(clip = "off") +
    theme(panel.grid.major.x=element_line(color = "gray", size=0.25),
                panel.grid.major.y=element_blank())

Diverging Bar chart

A diverging bar chart is a type of data visualization that displays values of a quantitative variable across two or more categories, where the bars are centered on a common baseline or a central axis. This type of chart is particularly useful for comparing differences in values that can have both positive and negative magnitudes, or when illustrating a relationship between two opposing factors or sentiments.

In a diverging bar chart, the bars extend in opposite directions from the central axis, creating a symmetrical pattern. This layout makes it easy to see the differences between categories and to identify patterns and trends within the data.

When choosing a diverging bar chart, make sure it is appropriate for your dataset and the type of information you want to convey. It is important to note that diverging bar charts may not be suitable for all types of data, especially if the differences between categories are not symmetrical or if the data has no clear central point. In such cases, other visualization methods, like standard bar charts or line charts, might be more suitable.

example data (positive and negative values)

Show the code

change_in_chop_bars_df <- tribble(~region_number, ~region, ~change,
                            1,"Western", -.35, 
                            2, "Central", -.26, 
                            3, "Greater Accra", .67,
                            4, "Volta", -.18, 
                            5, "Eastern", .47, 
                            6, "Ashanti",.56, 
                            7, "Western North", -.2,
                            8,  "Ahafo", .61, 
                            9, "Bono",.12,
                            10,"Bono East", 0.01,
                            11,"Oti",.07,
                            12,"Northern", .11,
                            13,"Savannah", .12,
                            14,"North East",-.31,
                            15,"Upper East",0,
                            16, "Upper West", .265)

change_in_chop_bars_df %>% 
    kable() %>%
    kable_styling() %>% 
  scroll_box(height = "400px")

region_number	region	change
1	Western	-0.350
2	Central	-0.260
3	Greater Accra	0.670
4	Volta	-0.180
5	Eastern	0.470
6	Ashanti	0.560
7	Western North	-0.200
8	Ahafo	0.610
9	Bono	0.120
10	Bono East	0.010
11	Oti	0.070
12	Northern	0.110
13	Savannah	0.120
14	North East	-0.310
15	Upper East	0.000
16	Upper West	0.265

Show the code

change_in_chop_bars_df %>% 
    mutate(pos_neg = ifelse(change >0, "positive", "negative")) %>% 
    ggplot(aes(x = reorder(region, change),
                         y = change,
                         fill =pos_neg)) +
  geom_col() +
  gssthemes() +
        scale_fill_manual(values = c(positive = positive_color,
                                                             negative = negative_color))+ 
  scale_y_continuous(labels = scales::percent) +
    labs(x = NULL,
             y = "change in number\nof chop bars")+
  coord_flip(clip = "off") +
    theme(panel.grid.major.x=element_line(color = "gray", size=0.25),
                panel.grid.major.y=element_blank(),
                axis.text.x = element_text(vjust = 0.5),
                legend.position = "none") +
      geom_hline(yintercept = 0, linetype = "dashed", color = "black")

Bar chart with hidden labels

Sometimes, you might want to hide certain text labels. For example, if the value is too small. In the example below, any value under 12% is hidden.

Show the code

n_chopbars_df_longformat %>%
    group_by(region) %>%
    mutate(percentage = number_of_chop_bars / sum(number_of_chop_bars)) %>%
    ungroup() %>%
    ggplot(aes(x = region,  y = percentage, fill = locality)) +
    geom_col() +
    coord_flip(expand = FALSE,
                         clip = "off") +
    scale_y_continuous(labels = scales::percent) +
    geom_text(aes(
        label = ifelse(percentage < .12, "",
                                     paste0(format(round(percentage * 100, 1), nsmall =1), "%")),
        y = ifelse(locality == "urban", 0.04, 0.91)
    ),
    size = 3.5,
    position = "stack") +
    gssthemes() +
    labs(x = NULL,
             y = "share of chop bars") +
    scale_fill_manual(
        values = c(urban = urban_color,
                             rural = rural_color),
        labels = c("Rural", "Urban")
    ) +
    guides(fill = guide_legend(reverse = TRUE)) +
    theme(
        panel.grid.major.x = element_line(color = "gray", size = 0.25),
        panel.grid.major.y = element_blank(),
        axis.text.x = element_text(vjust = 0.5)
    )

Filled bar chart with from positive to negative

You can use a filled bar chart to show relative shares within a group. This example shows the percentage of different opininions.

example data (sentiment data)

Show the code

sentiment_chop_bars <- tribble(~location, ~ `very positive`, ~ `positive`, ~  `neutral`,  ~ `negative`, ~"very negative",
                "urban", 15.2, 17.3, 50.2, 8, 9.3,
                "rural" , 25.2,34.2,31.2,6,3.4,
                "national", 20.2,25.75, 40.7,7,6.35)

sentiment_chop_bars %>% 
        kable() %>%
    kable_styling() 

location	very positive	positive	neutral	negative	very negative
urban	15.2	17.30	50.2	8	9.30
rural	25.2	34.20	31.2	6	3.40
national	20.2	25.75	40.7	7	6.35

Show the code

sentiment_chop_bars %>%
    pivot_longer(-location, names_to = "sentiment") %>%
    mutate(percent = value / 100) %>%
    # make sure the sentiment labels are factors so that they can be ordered correctly
    mutate(sentiment = factor(
        sentiment,
        levels = c(
            "very positive",
            "positive",
            "neutral",
            "negative",
            "very negative"
        )
    )) %>%
    ggplot(aes(x = location, y = percent, fill = sentiment)) +
    geom_col(width = 0.8, position = "stack") +
    gssthemes() +
    scale_y_continuous(labels = scales::percent,
                                         expand = c(0, 0)) +
    labs(x = NULL,
             y = "percentage",
             fill = "opinion on chop bars") +
    coord_flip(clip = "off") +
    scale_fill_manual(values = posneutralneg_color_scheme) +
    theme(
        panel.grid.major.x = element_line(color = "gray", size = 0.25),
        panel.grid.major.y = element_blank(),
        legend.position = "top"
    ) +
    guides(fill = guide_legend(
        nrow = 2,
        byrow = TRUE,
        title.position = "top"
    ))

multiple plots next to each other using facet_wrap()

R’s faceting system is a powerful way to make “small multiples”.

Show the code

n_chopbars_df_longformat %>% 
  ggplot(mapping = aes(x = reorder(region,number_of_chop_bars), fill = locality ,  y = number_of_chop_bars)) +
  geom_col(width = 0.8, position = "dodge") +
    geom_text(mapping = aes(label = scales::comma(number_of_chop_bars)), hjust = -0.2, 
                        position = position_dodge(width = 0.8)) +

  gssthemes() +
  scale_y_continuous( expand =c(0,NA),
                                        limits = c(0, 7000),
                                        labels = scales::comma,
                      breaks = scales::extended_breaks(only.loose = TRUE)) +
    labs(x = NULL,
             y = "number of chop bars")+
  coord_flip(clip = "off") +
    scale_fill_manual(values = c(urban = urban_color,
                                                             rural = rural_color))+
    theme(panel.grid.major.x=element_line(color = "gray", size=0.25),
                panel.grid.major.y=element_blank(),
                legend.position = "none") +
    # use the facet_wrap function to make small multiples
    facet_wrap(~locality)

multiple plots next to each other with separate bar legends

Sometimes using facet_wrap is not sufficient, especially if you want to add different legends to sub-plots, but still want to align different axis of plots into a single plots. To do this you can create 3 seperate plots and merge them using the patchwork package

example data (long format with quarters)

Show the code

n_chopbars_df_region_quarters  <- tribble(~region_number, ~region, ~national, ~urban, ~rural, ~quarter,
1,"Western", 567, 300, 267, "Q1",
2, "Central", 907, 600, 307,"Q1",
3, "Greater Accra", 6990, 4792, 2198, "Q1",
4, "Volta", 278, 78, 200, "Q1",
5, "Eastern", 787, 287, 500,"Q1",
6, "Ashanti",5621, 621, 5000,"Q1",
7, "Western North", 345, 145, 200,"Q1",
8,  "Ahafo", 678, 339, 339,"Q1",
9, "Bono",1268, 600, 668, "Q1",
10,"Bono East", 1268,1048, 220,"Q1",
11,"Oti",773, 500, 273, "Q1",
12,"Northern", 112, 12, 100,"Q1",
13,"Savannah", 1289,503, 786, "Q1",
14,"North East",378, 142, 236,"Q1",
15,"Upper East",1279,326, 953, "Q1",
16, "Upper West", 3782, 3456, 326, "Q1",
1,"Western", 602, 302, 300, "Q2",
2, "Central", 1000, 600, 400,"Q2",
3, "Greater Accra", 7990, 4792, 3198, "Q2",
4, "Volta", 346, 93, 253, "Q2",
5, "Eastern", 887, 387, 500,"Q2",
6, "Ashanti",6000, 2000, 4000,"Q2",
7, "Western North", 445, 195, 250,"Q2",
8,  "Ahafo", 701, 339, 362,"Q2",
9, "Bono",1444, 722, 722, "Q2",
10,"Bono East", 1368,1048, 290,"Q2",
11,"Oti",912, 534, 378, "Q2",
12,"Northern", 122, 17, 105,"Q2",
13,"Savannah", 1289,503, 786, "Q2",
14,"North East",534, 182, 352,"Q2",
15,"Upper East",1510,400, 1110, "Q2",
16, "Upper West", 3991, 3496, 495, "Q2")


n_chopbars_df_region_quarters %>% 
    kable() %>%
    kable_styling() %>% 
  scroll_box(height = "400px")

region_number	region	national	urban	rural	quarter
1	Western	567	300	267	Q1
2	Central	907	600	307	Q1
3	Greater Accra	6990	4792	2198	Q1
4	Volta	278	78	200	Q1
5	Eastern	787	287	500	Q1
6	Ashanti	5621	621	5000	Q1
7	Western North	345	145	200	Q1
8	Ahafo	678	339	339	Q1
9	Bono	1268	600	668	Q1
10	Bono East	1268	1048	220	Q1
11	Oti	773	500	273	Q1
12	Northern	112	12	100	Q1
13	Savannah	1289	503	786	Q1
14	North East	378	142	236	Q1
15	Upper East	1279	326	953	Q1
16	Upper West	3782	3456	326	Q1
1	Western	602	302	300	Q2
2	Central	1000	600	400	Q2
3	Greater Accra	7990	4792	3198	Q2
4	Volta	346	93	253	Q2
5	Eastern	887	387	500	Q2
6	Ashanti	6000	2000	4000	Q2
7	Western North	445	195	250	Q2
8	Ahafo	701	339	362	Q2
9	Bono	1444	722	722	Q2
10	Bono East	1368	1048	290	Q2
11	Oti	912	534	378	Q2
12	Northern	122	17	105	Q2
13	Savannah	1289	503	786	Q2
14	North East	534	182	352	Q2
15	Upper East	1510	400	1110	Q2
16	Upper West	3991	3496	495	Q2

Show the code

n_chopbars_df_region_quarters_longformat <- n_chopbars_df_region_quarters %>% 
    pivot_longer(-c(region_number, region, quarter),
                             names_to = "locality",
                             values_to = "number_of_chop_bars"
                             )


p_national <- n_chopbars_df_region_quarters_longformat %>% 
    filter(locality == "national") %>% 
    ggplot(mapping = aes(x     = reorder(region,desc(region_number)),
                                             y     = number_of_chop_bars)) +
    geom_col(width = 0.8,
                     aes( alpha = reorder(quarter, desc(quarter))),
                     position =  position_dodge(width = 0.8),
                     fill = national_color)+
    geom_text(mapping = aes(label = scales::comma(number_of_chop_bars),
                                                    group = reorder(quarter, desc(quarter))),
                        hjust = -0.2,
                        position = position_dodge(width = 0.8)
    ) +
    gssthemes() +
    scale_y_continuous( limits = c(0, 10000),
                                            labels = scales::comma) +
    scale_alpha_manual(values = seq(
        from = 0.5,
        to = 1,
        length.out = 2
    ),
    guide = guide_legend(reverse = TRUE))+
    labs(x = NULL,
             alpha = NULL,
             y = "number of chop bars",
             title = "National")+
    coord_flip(clip = "off") +
    theme(panel.grid.major.x = element_line(color = "gray", size=0.25),
                panel.grid.major.y = element_blank(),
                legend.position = "top",
                axis.text.x = element_text(vjust = 0.5),
                plot.title=element_text(hjust=0.5),
        plot.subtitle=element_text(hjust=0.5))



p_urban <- n_chopbars_df_region_quarters_longformat %>% 
    filter(locality == "urban") %>% 
    ggplot(mapping = aes(x     = reorder(region,desc(region_number)),
                                             y     = number_of_chop_bars,
                                             group = reorder(quarter, desc(quarter)))) +
    geom_col(width = 0.8,
                     aes( alpha = reorder(quarter, desc(quarter))),
                     position = "dodge",
                     fill = urban_color) +
    geom_text(mapping = aes(label = scales::comma(number_of_chop_bars)), hjust = -0.2, 
                        position = position_dodge(width = 0.8)) +
    gssthemes() +
    scale_y_continuous( limits = c(0, 10000),
                                            labels = scales::comma) +
    scale_alpha_manual(values = seq(
        from = 0.5,
        to = 1,
        length.out = 2
    ),
    guide = guide_legend(reverse = TRUE))+
    labs(x = NULL,
             alpha = NULL,
             y = "number of chop bars",
             title = "Urban")+
    coord_flip(clip = "off") +
  # move the title text and subtitle text to the middle
  theme(plot.title=element_text(hjust=0.5),
        plot.subtitle=element_text(hjust=0.5),
            panel.grid.major.x = element_line(color = "gray", size=0.25),
                panel.grid.major.y = element_blank(),
                axis.ticks.y= element_blank(),              
                legend.position = "top",
                axis.text.y = element_blank())

p_rural <- n_chopbars_df_region_quarters_longformat %>% 
    filter(locality == "rural") %>% 
    ggplot(mapping = aes(x     = reorder(region,desc(region_number)),
                                             y     = number_of_chop_bars,
                                             group = reorder(quarter, desc(quarter)))) +
    geom_col(width = 0.8,
                     aes( alpha = reorder(quarter, desc(quarter))),
                     position = "dodge",
                     fill = rural_color) +
    geom_text(mapping = aes(label = scales::comma(number_of_chop_bars)), hjust = -0.2, 
                        position = position_dodge(width = 0.8)) +
    gssthemes() +
    scale_y_continuous( limits = c(0, 10000),
                                            labels = scales::comma) +
    scale_alpha_manual(values = seq(
        from = 0.5,
        to = 1,
        length.out = 2
    ),
    guide = guide_legend(reverse = TRUE))+
    labs(x = NULL,
             alpha = NULL,
             y = "number of chop bars",
             title = "Rural")+
    coord_flip(clip = "off") +
  # move the title text and subtitle text to the middle
  theme(plot.title=element_text(hjust=0.5),
        plot.subtitle=element_text(hjust=0.5),
            panel.grid.major.x = element_line(color = "gray", size=0.25),
                panel.grid.major.y = element_blank(),
                axis.ticks.y= element_blank(),              
                legend.position = "top",
                axis.text.y = element_blank())


p_national + p_urban + p_rural

bar chart with a singe category hightlighted on top

with a bit of effort, a variable can also be shown on its own. In this case you can use facet_grid()

Show the code

n_chopbars_df_with_national   %>% 
    mutate(highlight = region == "Ghana") %>% 
    # Create a ggplot object and set the aesthetic mappings
    # use the reorder function to reorder the the bars
    ggplot(mapping = aes(x = reorder(region,number_of_chop_bars), 
                                             y = number_of_chop_bars,
                                             fill = highlight)) +
    geom_col(width = 0.8) +
    geom_text(mapping = aes(label = scales::comma(number_of_chop_bars)), hjust = -0.2) +
    gssthemes() +
    scale_y_continuous( expand = c( 0, 0 ),
                                            limits = nicelimits,
                                            labels = scales::comma,
                                            breaks = scales::extended_breaks(only.loose = TRUE)) +
    scale_x_discrete(labels= function(x) highlight(x, "Ghana", "black")) +
    scale_fill_manual(values = c(`FALSE` = national_color, 
                                                             `TRUE` = scales::muted(national_color)),
                                        guide = "none") +
    labs(x = NULL,
             y = "number of chop bars" )+
    coord_flip(clip = "off") +
    theme(panel.grid.major.x = element_line(color = "gray", size=0.25),
                panel.grid.major.y = element_blank(),
                axis.text.x = element_text(vjust = 0.5),
                strip.text = element_blank(),
                axis.text.y = element_markdown()) +
    facet_grid(fct_rev(as.factor(highlight))~.,scales = "free",
                         space = "free_y")# to make this work you need to make the y axis a markdown format

Population pyramid

An illustration of the age and gender distribution of a population is called a population pyramid. It is widely used to illustrate both the current trend in population growth and a population’s makeup. In most cases, the chart is divided into two equal parts, one for males and the other for women. The age groups are frequently displayed along a vertical axis, with the youngest age group at the bottom and the oldest at the top. To make sure the ggplot organises that age categorically, instead of alphabetically, you need to make sure you make the age grouping a factor variable. You can use scale_y_continuous(labels = abs) to make the variables at the base absolute values.

example data (Population pyramid)

Show the code

# Sample data
age_pyramid_df <- data.frame(
  age_group = rep(c("0-4", "5-9", "10-14", "15-19", "20-24", "25-29", "30-34", "35-39", "40-44", "45-49", "50-54", "55-59", "60-64", "65-69", "70-74", "75+"), 2),
  gender = rep(c("Male", "Female"), each = 16),
  count = c(100, 95, 90, 85, 80, 75, 70, 65, 60, 55, 50, 45, 40, 35, 30, 25,
            105, 100, 95, 90, 85, 80, 75, 70, 65, 60, 55, 50, 45, 40, 35, 30)
) %>% 
    mutate(age_group  = factor(age_group, levels = c("0-4", "5-9", "10-14", "15-19", "20-24", "25-29", "30-34", "35-39", "40-44", "45-49", "50-54", "55-59", "60-64", "65-69", "70-74", "75+")))


age_pyramid_df %>% 
    kable() %>%
    kable_styling() %>% 
  scroll_box(height = "400px")

age_group	gender	count
0-4	Male	100
5-9	Male	95
10-14	Male	90
15-19	Male	85
20-24	Male	80
25-29	Male	75
30-34	Male	70
35-39	Male	65
40-44	Male	60
45-49	Male	55
50-54	Male	50
55-59	Male	45
60-64	Male	40
65-69	Male	35
70-74	Male	30
75+	Male	25
0-4	Female	105
5-9	Female	100
10-14	Female	95
15-19	Female	90
20-24	Female	85
25-29	Female	80
30-34	Female	75
35-39	Female	70
40-44	Female	65
45-49	Female	60
50-54	Female	55
55-59	Female	50
60-64	Female	45
65-69	Female	40
70-74	Female	35
75+	Female	30

Show the code

age_pyramid_df %>% 
    ggplot() +
  geom_bar(aes(x = age_group, 
                         y = ifelse(gender == "Male", -count, count), fill = gender), stat = "identity") +
  scale_y_continuous(labels = abs) +
    scale_fill_manual(values = c("Male" = male_color,
                                                            "Female" = female_color)) +
  coord_flip() +
    labs(x = "age group", y = "count") +
    gssthemes() 

Line Charts

A line chart is a graphical representation used to display continuous data over a period of time or across one or multiple categories. A user should use a line chart when they want to show trends, patterns, or changes over time, especially for a large number of data points or continuous data.

example data (line)

Show the code

line_chart_df <-tribble(~year, ~elder,
                                            1960, 213477, 
                                            1984, 493359,
                                            2000, 1363410, 
                                            2010, 1640220, 
                                            2021, 1993300) %>% 
    mutate(year = factor(year))

line_chart_df %>% 
    kable() %>%
    kable_styling() 

year	elder
1960	213477
1984	493359
2000	1363410
2010	1640220
2021	1993300

Show the code

line_chart_df %>% 
    mutate(year = as.numeric(levels(year))) %>% 
    ggplot(aes(x = year, y = elder)) +
    geom_line(color = national_color, linewidth = 1.5) +
    geom_point(color = national_color, size = 3)+
    scale_x_continuous(breaks = as.numeric(levels(line_chart_df$year))) +
    gssthemes()

Line chart multiple categories

example data (Line Chart with 2 categories)

Show the code

line_chart_df_by_sex <-tribble(~year, ~Male, ~Female,
                                            1960,4572, 2562,
                                            1984,5572, 4572,
                                            2000,6787, 5887,
                                            2010,7211, 6500,
                                            2021, 8000, 7900) %>% 
    pivot_longer(-c(year), 
                             names_to = "sex",
                             values_to = "number")

line_chart_df_by_sex %>% 
    kable() %>%
    kable_styling() %>% 
  scroll_box(height = "400px")

year	sex	number
1960	Male	4572
1960	Female	2562
1984	Male	5572
1984	Female	4572
2000	Male	6787
2000	Female	5887
2010	Male	7211
2010	Female	6500
2021	Male	8000
2021	Female	7900

Show the code

line_chart_df_by_sex %>%
    ggplot(aes(x = year, y = number, color = sex )) +
    geom_line(linewidth = 1.5) +
    geom_point(size = 3) +
    scale_y_continuous() + 
    scale_color_manual(values = c("Male" = male_color,
                                                             "Female" = female_color))  +
        scale_x_continuous(breaks = as.numeric(levels(line_chart_df$year))) +
        scale_y_continuous( expand = c( 0, 0 ),
                                            limits = nicelimits,
                                            labels = scales::comma,
                                            breaks = scales::extended_breaks(only.loose = TRUE)) +

    gssthemes() +
    theme(legend.key = element_blank())

Line chart multiple categories with labels

To add value labels to a line chart, you can use the geom_text() function or the geom_text_repel() function from the ggrepel package if you want to make sure the labels do not overlap. You can use show.legend=FALSE to make sure the legend does not contain text.

Show the code

line_chart_df_by_sex %>%
    ggplot(aes(x = year, y = number, color = sex)) +
    geom_line(linewidth = 1.5) +
    geom_point(size = 3) +
    geom_text(
        aes(label = number),
        vjust = ifelse(line_chart_df_by_sex$sex == "Male",-1, 2),
        show.legend = FALSE
    ) +
    scale_color_manual(values = c("Male" = male_color,
                                                                "Female" = female_color))  +
    scale_x_continuous(breaks = as.numeric(levels(line_chart_df$year)),
                                         limits = c(1960, 2030)) +
    scale_y_continuous(
        expand = c(0 , 0),
        limits = c(0, 9000),
        labels = scales::comma,
        breaks = scales::extended_breaks(only.loose = TRUE)
    ) +
    gssthemes() +
    theme(legend.key = element_blank())

Line chart with group labels showing at end of line

Instead of a legend, you can also add the line labels to the end of the lines

Show the code

library(ggrepel)
line_chart_df_by_sex %>%
    # change the sex label so that it only shows up at the end
    # by making it NA for all other years
    mutate(sex_label = ifelse(year == max(year), as.character(sex), NA_character_)) %>% 
    ggplot(aes(x = year, y = number, color = sex)) +
    geom_line(linewidth = 1.5) +
    geom_point(size = 3) +
  geom_text_repel(aes(label = sex_label), hjust = -0.2,
                                  xlim = c(max(line_chart_df_by_sex$year) +5, NA),
                                direction = "y",
                                    hjust         = 0,
                                       segment.curvature = -1e-20,
                                segment.angle = 90,
                                  min.segment.length = 0.2,
                                        show.legend = FALSE) +
    scale_color_manual(values = c("Male" = male_color,
                                                                "Female" = female_color))  +
    scale_x_continuous(breaks = unique(line_chart_df_by_sex$year),
                                         limits = c(1960, max(line_chart_df_by_sex$year) + 20)) +
    scale_y_continuous(
        expand = c(0 , 0),
        limits = c(0, 9000),
        labels = scales::comma,
        breaks = scales::extended_breaks(only.loose = TRUE)
        ) +
    gssthemes() +
    theme(legend.position = "none")

step chart

A step chart is a suitable choice over a line chart when you want to represent data that changes at specific intervals and remains constant between those intervals. It is particularly useful for visualizing data with discrete jumps or where the value holds steady until the next change occurs.

Step charts are often used in scenarios where data has a clear start and end point for each period, such as price changes, tax rate adjustments, or policy changes. By using a step chart, you can emphasize the moments when the changes happen and highlight the periods of stability between them, making it easier to identify and analyze the impacts of these changes. In contrast, a line chart would connect the data points directly, which might imply a continuous transition between values, potentially misrepresenting the nature of the data.

Show the code

line_chart_df %>% 
    mutate(year = as.numeric(levels(year))) %>% 
    ggplot(aes(x = year, y = elder)) +
    geom_step(color = national_color, linewidth = 1.5) +
    scale_x_continuous(breaks = as.numeric(levels(line_chart_df$year)))  +
      gssthemes() +
  scale_y_continuous( expand = c( 0, 0 ),
                      limits = c(0,2500000),
                      breaks = scales::extended_breaks(only.loose = TRUE),
                                        labels = scales::comma)

Smoothed line chart

In some cases, you might want to use a smoothed line (that still goes through all the points). In this case you can use geom_xspline(spline_shape = -0.5) from the ggalt package

Show the code

library(ggalt)
line_chart_df %>% 
    mutate(year = as.numeric(levels(year))) %>% 
    ggplot(aes(x = year, y = elder)) +
    geom_point(color = national_color, size =3)+ 
    ggalt::geom_xspline(color = national_color, size = 1.5, spline_shape = -0.5) +
    scale_x_continuous(breaks = as.numeric(levels(line_chart_df$year))) +
    gssthemes()

lollipop chart

A lollipop chart is used visually represent time series data where magnitude of occurrence is to be communicated. It shows the relationship between a numeric and a categorical variable, as well as identifying trends or patterns over time. They can be used to display numerical data such as quantities and magnitude. To make sure that the data is shown above the right year, you can use as.factor() to make the numeric value a factor

Show the code

# Create lollipop plot
line_chart_df %>%
    ggplot(mapping=aes(x =year, y = elder)) + 
  geom_segment(aes(xend = year, yend = 0), linewidth = 1) +
  geom_point(aes(size = elder),colour=national_color) +
  geom_text(aes(label =scales::comma(elder)), vjust= ifelse(line_chart_df$elder<500000, -1.5,-4))+
  scale_size(range=c(2,24)) +
  labs(x = "Census Year", y = "Population") +
  gssthemes() +
  theme(legend.position = "none") +
  scale_y_continuous( expand = c( 0, 0 ),
                                        # due to bubble size scale limit has to be manually set
                      limits = c(0,2500000),
                      breaks = scales::extended_breaks(only.loose = TRUE),
                                        labels = scales::comma)

Area charts

Area chart (single category)

An area chart is a graphical representation used to display quantitative data over a period of time or across different categories. A user should use an area chart when they want to emphasize the magnitude of change or cumulative totals, visualize the composition of data, or show trends across multiple series while highlighting their differences.

Show the code

line_chart_df %>% 
    mutate(year = as.numeric(levels(year))) %>% 
    ggplot(aes(x = year, y = elder)) +
    geom_line(color = national_color, linewidth = 1.5) +
    geom_area(fill = national_color, alpha = .5)+
    scale_x_continuous(breaks = as.numeric(levels(line_chart_df$year))) +
    gssthemes()

Area chart (stacked)

Show the code

line_chart_df_by_sex %>% 
    ggplot(aes(x=year, y=number, fill=sex)) + 
    geom_area() +
    scale_fill_manual(values = c("Male" = male_color,
                                                             "Female" = female_color)) +
    scale_y_continuous( expand = c( 0, 0 ),
                                            # due to bubble size scale limit has to be manually set
                                            breaks = scales::extended_breaks(only.loose = TRUE),
                                            labels = scales::comma,
                                            limits = nicelimits) +
    gssthemes()

Area chart (Proportional)

In a proportional stacked area chart, the total for each year consistently amounts to one hundred percent, with the values of individual groups expressed as percentages. First you need to compute these relative percentages for each year

Show the code

line_chart_df_by_sex  %>%
  group_by(year ) %>%
  mutate(percentage = number / sum(number)) %>% 
    ggplot(aes(x=year, y= percentage, fill=sex)) + 
    geom_area(alpha = .7) +
    scale_fill_manual(values = c("Male" = male_color,
                                                             "Female" = female_color)) +
    scale_y_continuous( expand = c( 0, 0 ),
                                            # due to bubble size scale limit has to be manually set
                                            breaks = scales::extended_breaks(only.loose = TRUE),
                                            labels = scales::percent,
                                            limits = nicelimits) +
    gssthemes()

Ribbon chart

One can use a ribbon chart to put some extra attention to the difference between two lines. To do so, you will have to create two dataframes, one in long format and one in wide format.

Show the code

# Create the data frame
area_df <- tibble(
    year = c(2010, 2011, 2012, 2013, 2014, 2015),
    `rural growth` = c(2.1, 2.5, 2.8, 3.2, 3.5, 3.9),
    `urban growth` = c(2.8, 3.1, 3.4, 3.7, 4.0, 4.3)
)

Show the code

area_df_long <- area_df %>%
    pivot_longer(-year)
# Create the plot
ggplot() +
    geom_line(data = area_df_long,
                        aes(x = year, y = value, col = name),
                        linewidth = 2) +
    geom_ribbon(
        data = area_df,
        aes(x = year,
                ymin = `rural growth`,
                ymax =  `urban growth`),
        alpha = 0.2,
        fill = national_color
    ) +
        scale_color_manual(values = c( `urban growth` = urban_color,
                                                             `rural growth` = rural_color)) +
    labs(color = NULL, x = "Year", y = "Growth (%)") +
    gssthemes() +
    theme(legend.position = "bottom")

Density Chart

A density plot, also known as a kernel density plot, is a sort of graphical data representation that depicts data distribution over a continuous interval or range of values. It is similar to a histogram in that it displays the estimated probability density function of the data rather than the frequency of data points in each bin.

In data analysis and statistics, density plots are often used to show the shape of the data distribution, including information about the mode, skewness, and outliers. They are especially effective for spotting trends in data that other types of plots, such as scatterplots or boxplots, may miss.

example data (density and histogram)

Show the code

# Create a data frame with three columns for agriculture, industry, and services
densitydf <- tibble(agriculture = rnorm(1000, mean = 50, sd = 10),
                   industry = rnorm(1000, mean = 75, sd = 15),
                   services = rnorm(1000, mean = 100, sd = 20))


# Convert the data frame to a long format
densitydf_long <- densitydf %>%
    pivot_longer(cols = everything(), 
                             names_to = "sector")

Show the code

# Create a density plot using geom_density with colors red, blue, and green
densitydf_long %>%
    ggplot(aes(x = value, fill = sector)) +
  geom_density(alpha = 0.5) +
  scale_fill_manual(values = c(industry = industry_color,
                                                         agriculture = agric_color,
                                                         services = services_color)) +
  labs(title = "Density Plot for Agriculture, Industry, and Services", x = "Value", y = "Density")+
    gssthemes()

Histogram

Like a density chart, a histogram is a graphical representation of the distribution of a dataset. It is an estimate of the probability distribution of a continuous variable (quantitative variable) or a discrete variable in certain cases. To construct a histogram, the data is divided into a set of intervals, also known as bins or buckets, and the number of data points falling into each bin is represented by the height of a bar. The bins are usually specified as consecutive, non-overlapping intervals of a variable.

Show the code

densitydf_long %>%
    ggplot(aes(x = value, fill = sector)) +
  geom_histogram(alpha = 0.5) +
  scale_fill_manual(values = c(industry = industry_color,
                                                         agriculture = agric_color,
                                                         services = services_color)) +
  labs(title = "Density Plot for Agriculture, Industry, and Services", 
         x = "Value",
         y = "Count")+
    gssthemes()

Histograms (dodged)

Instead of having the bars of the histogram stack you can also choose to have them next to each other.

Show the code

densitydf_long %>%
    ggplot(aes(x = value, fill = sector)) +
  geom_histogram(position = "dodge") +
  scale_fill_manual(values = c(industry = industry_color,
                                                         agriculture = agric_color,
                                                         services = services_color)) +
  labs(title = "Density Plot for Agriculture, Industry, and Services", 
         x = "Value",
         y = "Count")+
    gssthemes() +
    theme(legend.position = "bottom")

Box plot

Boxplots can be used to display the distribution of a numerical data. They are particularly useful for summarizing the spread and skewness of a distribution, as well as identifying outliers. The box in the plot represents the interquartile range (IQR), which spans from the first quartile (Q1) to the third quartile (Q3) of the data. The median is represented by a horizontal line inside the box. To make a boxplot, you can use the geom_boxplot() function.

Show the code

densitydf_long %>%
    ggplot(aes(x = sector, y = value, fill = sector)) +
  geom_boxplot() +
  scale_fill_manual(values = c(industry = industry_color,
                                                         agriculture = agric_color,
                                                         services = services_color)) +
  labs(title = "Boxplot for income in Agriculture, Industry, and Services", 
         x = NULL,
         y = "Income")+
    gssthemes() +
    theme(legend.position = "bottom")

Heatmap

Heatmap is a graphical representation of data that uses colour-coded cells to represent values. Heatmaps are commonly used to visualize the distribution and density of data points within a particular area or to show patterns and correlations in large datasets.

In a heatmap, each cell represents a specific data point or group of data points, and the colour of the cell indicates the value of the data. Heatmaps can be used to visualize changes in data over time, display geographic data, such as population density or weather patterns, highlight outliers or anomalies in a dataset, which may indicate errors or unusual pattern among others

Show the code

# Sample data : Changes in data overtime
heatmap_df <- tibble(
    Sector = c(
        "Agriculture",
        "Manufacturing",
        "Trade",
        "Transport",
        "Services"
    ),
    Q1 = c(3, 4, 5, 6, 7),
    Q2 = c(4, 5, 6, 7, 8),
    Q3 = c(5, 6, 7, 8, 9),
    Q4 = c(6, 7, 8, 9, 10)
) %>%
    pivot_longer(cols = -Sector,
                             names_to = "Quarter")

heatmap_df %>%
    kable() %>%
    kable_styling() %>%
    scroll_box(height = "400px")

Sector	Quarter	value
Agriculture	Q1	3
Agriculture	Q2	4
Agriculture	Q3	5
Agriculture	Q4	6
Manufacturing	Q1	4
Manufacturing	Q2	5
Manufacturing	Q3	6
Manufacturing	Q4	7
Trade	Q1	5
Trade	Q2	6
Trade	Q3	7
Trade	Q4	8
Transport	Q1	6
Transport	Q2	7
Transport	Q3	8
Transport	Q4	9
Services	Q1	7
Services	Q2	8
Services	Q3	9
Services	Q4	10

Show the code

# create heatmap using ggplot
heatmap_df %>% 
    ggplot(aes(x = Quarter, y = Sector, fill = value)) +
  geom_tile() +
  scale_fill_gradientn(colours = incidence_color_scheme) +
  theme_minimal() +
  labs(title = "Economic Activity by Sector", x = "Quarter", y = "Sector", fill = "Value") + 
    gssthemes()+
    theme(legend.position = "bottom")

Proportion Charts

Pie charts

A pie chart is a circular statistical picture that divides into slices to represent numerical quantities. Each pie slice symbolizes a certain category, and the size of each slice is inversely proportionate to the amount it represents. Pie charts are used to show how several categories are spread within a whole.

Pie charts are the most effective for comparing parts of an entire. They do not show changes over time or relationships between variables. When there are many categories or not many differences between the groupings, pie charts might be difficult to read and comprehend.

example data (Pie chart with 2 categories)

Show the code

# Sample data
df_pie <- data.frame(locality = c("Urban", "Rural"),
  deaths = c(4500, 2100))


df_pie %>% 
    kable() %>%
    kable_styling()

locality	deaths
Urban	4500
Rural	2100

Show the code

# Pie chart
df_pie %>% 
    ggplot() +
  geom_col(aes(x = "", y = deaths, fill = locality)) +
  coord_polar(theta = "y")+
    scale_fill_manual(values = c("Urban" = urban_color, "Rural"=  rural_color)) +
    gssthemes() +
    theme(axis.text = element_blank(),
                axis.ticks.y = element_blank(),
                axis.text.y= element_blank(),
                panel.grid.major.y = element_blank(),
                axis.text.x = element_blank(),
                panel.grid = element_blank()) +
    labs(x = NULL, y = NULL, fill = "Locality")

Pie chart with percentage value labels showing

Show the code

df_pie %>% 
    ggplot(aes(x = "",
                         y = deaths, 
                         fill = locality)) +
  geom_col() +
    coord_polar(theta = 'y') +
    scale_fill_manual(values = c("Urban" = urban_color, "Rural"=  rural_color)) +
    geom_text(aes(label = percent(deaths/sum(deaths, na.rm = T))),
                     color = "white", size = 8, position = position_stack(vjust = 0.5)) +
    gssthemes() +
    theme(axis.text = element_blank(),
                axis.ticks.y = element_blank(),
                axis.text.y= element_blank(),
                panel.grid.major.y = element_blank(),
                axis.text.x = element_blank(),
                panel.grid = element_blank())  +
    labs(x = NULL, y = NULL)

Pie chart with group labels showing on slice and with no legend

Show the code

df_pie %>% 
    ggplot(aes(x = "", y = deaths, fill = locality))+
    geom_col() +
    coord_polar(theta = 'y') +
    geom_text(aes(label = locality, 
                                color = "black"), 
                        size = 8,
                        position = position_stack(vjust = 0.5),
                        vjust = -1.5) +
    scale_fill_manual(values = c("Urban" = urban_color, "Rural"=  rural_color)) +
    geom_text(aes(label = percent(deaths/sum(deaths, na.rm = T))),
                     color = "white", size = 8, position = position_stack(vjust = 0.5)) +
    gssthemes() +
    theme(axis.text = element_blank(),
                axis.ticks.y = element_blank(),
                axis.text.y= element_blank(),
                panel.grid.major.y = element_blank(),
                axis.text.x = element_blank(),
                panel.grid = element_blank(),
                legend.position = "none")  +
    labs(x = NULL, y = NULL)

Show the code

# # Sample data
# df <- data.frame(
#   group = c("A", "B", "C", "D"),
#   value = c(40, 10, 15, 35)
# )
# 
# # Donut chart
# ggplot(df) +
#   geom_col(aes(x = "", y = value, fill = group)) +
#   coord_polar(theta = "y") +
#   theme_void() +
#   xlim(-1.5, 1.5) +
#   annotation_custom(grob = grid::rectGrob(gp = grid::gpar(col = NA, fill = "white")), xmin = -1, xmax = 1, ymin = -1, ymax = 1)+
#   gssthemes()

Waffle chart

A waffle chart, also known as a square pie chart or a mosaic plot, is a type of chart that displays parts of a whole for categorical quantities. It is similar to a pie chart but uses squares instead of wedges to represent the proportions. Each square in the chart represents a fixed quantity and the total number of squares represents the total quantity.

Waffle charts are used in similar situations as pie charts, to show the relative proportions of different categories within a whole. They can be useful when you want to compare parts of a whole or when you want to display multiple pie charts with the same scale in a small space. Waffle charts can also be easier to read and interpret than pie charts when there are many categories or if the differences between the categories are small. To make a waffle chart in R, you can use the waffle() function from the waffle package. This package works woth named vectors, so to use it you need to transform your tibble to a named vector.

example data (waffle chart)

Show the code

data_waffle_chart <- tribble(~Type ,~ Food, ~ number,
                                                        "Starch",  "Fufu",   52,
                                                        "Starch", "Banku",  18,
                                                        "Starch",  "Rice",  10,
                                                        "Fish", "Tilapia", 6,
                                                        "vegetables", "Tomato" , 2,
                                                        "vegetables", "Okra" , 2) 
data_waffle_chart %>% 
    kable() %>%
    kable_styling() 

Type	Food	number
Starch	Fufu	52
Starch	Banku	18
Starch	Rice	10
Fish	Tilapia	6
vegetables	Tomato	2
vegetables	Okra	2

Show the code

library(waffle)

vect_waffle_chart <- data_waffle_chart %>% pull(number)
names(vect_waffle_chart) <- data_waffle_chart %>% pull(Food)

waffle(vect_waffle_chart) + 
    gssthemes() + 
    theme(axis.text.x = element_blank(),
                axis.text.y = element_blank(),
                axis.ticks = element_blank(),
                legend.position = "bottom") +
    scale_fill_manual(values = neutral_color_scheme)+
    labs(fill = NULL)

Treemap

A treemap is used when you want to visualize hierarchical data or display the composition of a whole using nested rectangles. It is particularly useful for representing large amounts of data with varying proportions in a compact and space-efficient manner. Each rectangle within the treemap represents a category or sub-category, with the size of the rectangle corresponding to the value or magnitude of that specific category. To make a treemap chart in R, you can use the geom_treemap() function from the treemapify package.

Show the code

library(treemapify)

data_waffle_chart %>% 
    ggplot(aes(area = number, fill = Type, label = Food, subgroup = Type)) +
    geom_treemap(colour = "white") +
    geom_treemap_text(grow = T, reflow = T, colour = "white",  family = GSS_font) +
    geom_treemap_subgroup_border( colour = "white") +
    scale_fill_brewer(palette = "Set1") + 
    gssthemes() + 
    theme(axis.text.x = element_blank(),
                axis.text.y = element_blank(),
                legend.position = "bottom") +
    scale_fill_manual(values = neutral_color_scheme[c(1,5, 3)])+
    labs(fill = "type of food")

Dumbell chart

also known as connected dot plot, it shows the changes in a variable between two different conditions or points in time. It is frequently employed to contrast two connected variables across various categories. Each dot on the graph, which is made up of two dots connected by a line or bar, reflects the value of one of the variables for a certain category.

Dumbbell charts are helpful for comparing the range of a variable across various groups or for displaying change over time for multiple groups. To make a waffle chart in R, you can use the geom_dumbell() function from the ggalt package.

Show the code

library(ggalt)

n_chopbars_df %>%
    ggplot() +
    geom_dumbbell(aes(y = region, x = rural, xend = urban),
        size = 1,
        color = national_color,
        size_xend = 5,
        size_x = 5,
        colour_x = rural_color,
        colour_xend = urban_color,
        dot_guide = FALSE,
        dot_guide_size = 0.25,
        show.legend = TRUE
    ) +
    labs(x = NULL, y = NULL)+
    gssthemes()  +
    geom_point(data = tibble(x = NA_integer_, y= NA_integer_, fill = c("urban", "rural")),
                         mapping = aes(x =x, y = y, color =fill)) +
    scale_color_manual(name = "", values = c("urban" = urban_color,
                                                                                        "rural" = rural_color) )  +
    theme(legend.key = element_rect(fill = NA),
                legend.position = "bottom")+
    guides(colour = guide_legend(override.aes = list(size=5)))

Sankey Chart

A Sankey chart (also known as a Sankey diagram) is a specific type of flow diagram, where the width of the bands is proportional to the flow quantity. This makes Sankey charts useful for visualizing a flow or transfer of some quantity between different time point. To make a Sankey chart in ggplot, you can use the ggsankey library.

Show the code

library(ggsankey)

sankey_df <- tribble(~Q1, ~Q2, ~Q3,
                     "Labour force", "Labour force", "Labour force",
                     "Labour force", "Labour force", "Outside Labour force",
                     "Labour force", "Outside Labour force", "Labour force",
                     "Outside Labour force", "Outside Labour force", "Outside Labour force",
                     "Outside Labour force", "Outside Labour force", "Labour force") %>%
  make_long(Q1, Q2, Q3)

sankey_df %>% 
    kable() %>%
    kable_styling() 

x	node	next_x	next_node
Q1	Labour force	Q2	Labour force
Q2	Labour force	Q3	Labour force
Q3	Labour force	NA	NA
Q1	Labour force	Q2	Labour force
Q2	Labour force	Q3	Outside Labour force
Q3	Outside Labour force	NA	NA
Q1	Labour force	Q2	Outside Labour force
Q2	Outside Labour force	Q3	Labour force
Q3	Labour force	NA	NA
Q1	Outside Labour force	Q2	Outside Labour force
Q2	Outside Labour force	Q3	Outside Labour force
Q3	Outside Labour force	NA	NA
Q1	Outside Labour force	Q2	Outside Labour force
Q2	Outside Labour force	Q3	Labour force
Q3	Labour force	NA	NA

Show the code

ggplot(
    sankey_df,
    aes(
        x = x,
        next_x = next_x,
        node = node,
        next_node = next_node,
        fill = factor(node),
        label = node
    )
) +
    geom_sankey(flow.alpha = 0.5,
                            node.color = 1,
                            show.legend = FALSE) +
    geom_sankey_label(
        size = 3.5,
        color = 1,
        fill = "white",
        family = "century gothic bold"
    ) +
    scale_fill_manual(
        values = c("Outside Labour force" = negative_color, "Labour force" = positive_color),
        labels =  c("Outside Labour force" = "Outside Labour force", "Labour force" =
                                    "Labour force")
    ) +
    scale_x_discrete(position = "top") +
    gssthemes() +
    theme(
        panel.grid.major.y = element_blank(),
        panel.grid.minor.y = element_blank(),
        axis.text.y = element_blank(),
        axis.ticks.y = element_blank(),
        axis.title.x = element_blank()
    )