Recreate a Du Bois Bar Chart

Last updated on 2026-02-04 | Edit this page

Estimated time: 16 minutes

Part of this work will invovle creative attempts at making changes to data visualizations. While the first few exercises are intended to help students become familiar with the basic steps, the final independent exercise is meant to introduce more gaps. If possible, emphasize to students that if they struggle on the final independent exercise it is not about making a perfect graph, but learning how to make changes on their own terms.

Overview

Questions

  • How can I read tabular data to plot a bar graph in R?
  • How can I use ggplot to organize and format a bar graph in R?
  • What are the characteristics of Du Bois data visualizations?
  • How can I maintain a reproducible record of my data visualizations?
  • How can I use color to improve accessibility of data visualizations?
  • How can I engage my own creativity in data visualizations?

Objectives

  • Create bar graph variations based on historical and contemporary tabular data.
  • Develop basic R code to create bar graphs.
  • Examine the characteristics of the data visualization from 1900 exhibitions.
  • Use transparent, legible, and shareable record of how you created your own unique data visualizations.
  • Create an exported HTML file of your completed Notebook that your instructor may ask you to submit.
  • Differentiate color usage to improve legibility of findings from tabular data.
  • Implement your own creative ideas through data visualization.

Black Literacy After Emancipation

Callout

This interactive exercise is inspired by the annual #DuBoisChallenge. The #DuBoisChallenge is a call to scientists, students, and community members to recreate, adapt, and share on social media the data visualzations created by W.E.B. Du Bois and his collaborators in 1900. Before doing the interactive exercise, please read this article about the Du Bois Challenge: https://nightingaledvs.com/the-dubois-challenge/

Why use R to create graphics?

Data can usually be presented in a table or in a tabular format. You have probably seen tabular data in a spreadsheet software like Excel and to interact with the tabular data, you can click directly onto the table. While graphic user interface (GUI) programs such as Excel are commonly used, they can be limited in their user output. GUIs are often limited to a set of style parameters for creating graphics. Coding programs like R have more capabilities including graphics. In R, a data frame is is a data type that stores/holds tabular data. The first step to creating data visualization in R is to read tabular data into an R dataframe. This is like double clicking a file to open it in other computer programs like in Excel. But with R, we use code.

The R code to do this uses an <- arrow pointed at the name of the data frame and a read.csv function command followed by the web address within paraentheses where a csv (comma separated values) data file is located.

Presentation of the historical cross-national literacy data

Du Bois’ created many data portraits to share the story of Black Americans post-emancipation. The “Black Literacy After Emancipation” graphic at the top of this page shows mass education as one important strategy for furthering and deepining emancipation for Black Americans and others. In this workbook, you will recreate two bar plots: 1) Du Bois’ visualization of Black illiteracy rates in the US compared to illiteracy rates in other countries using 1900 data; and 2) Du Bois’ visualization using data on Black college attainment in the US today.

An important context of Du Bois’s graph of Black illiteracy is that literacy was illegal for enslaved people in the U.S. until emancipation and the Confederacy’s defeat during the Civil War. Illiteracy then declined rapidly as Black Americans sought to empower themselves through education. The Du Bois plotted this decline in illiteracy among Black residents in the state of Georgia in the figure below. They used decennial US census illiteracy rates for Georgia from 1860 to 1890 that are available here. They likely wrote “50%?” for the 1900 illiteracy rate because the Census did not publish 1900 illiteracy rates (available here) until several months after the Paris Exposition. For this exercise, we’re going to read in data from a website. And we’re going to place the data into a dataframe named d_literacy_country and d_college_country.

d_literacy_country | column_name | Description | |—|—|—| | country | Name of country | | illiteracy | illiteracy percentage |

d_college_country | column_name | Description | |—|—|—| | country | Name of country | | college | college percentage |

How can I input tabular data to plot a bar graph in R?

The first step for data visualization in R is to read data into an R dataframe. This is like double clicking a file to open it in other computer programs. But with R, we use code.

For this exercise, we’re going to read in data from a website. And we’re going to place the data into a dataframe named d_literacy_country.

There is no record of the exact data used by the Du Bois team for this bar graph. And the Du Bois graph curiously does not include tick marks with a labeled axis scale to show what exact values each bar represents. Why? Perhaps the Du Bois team wanted to emphasize that the bar graph was a rough comparison of illiteracy rates because of varied timing, methods, and national boundaries for measuring illiteracy rates at the time. The length of the “Negroes U.S.A” bar likely represents the national Black illiteracy rate of 57.1% reported by the 1890 US Census (see reported “Russie” (Russia) bar correspond to the national US Black illiteracy rate in the 1890 US Census (see here). So our data derives illiteracy rates for other countries based on the length of each country’s bar relative to “Negroes U.S.A.” bar, presuming the “Negroes U.S.A.” bar represents 57.1%.

The R code to read in this data this uses an <- arrow pointed at the name of the data frame and a read.csv function command followed by the web address within parentheses where a csv (comma separated values) data file is located. It looks like this:

d_literacy_country <- read.csv("web_address_with_data/data_file_name.csv") After writing this code, we can write the name of the data frame

d_literacy_country again on a separate line. This will list all of the data in the data frame.

Challenge

Challenge 1: Reading the data

If everything is working correctly, you should be given a table. Here is the code executed unsuccessfully because the blank.

d_literacy_country <- _____.csv("https://raw.githubusercontent.com/HigherEdData/Du-Bois-STEM/refs/heads/main/data/d_literacy_country.csv")

d_literacy_country

Show me the solution

d_literacy_country <- read.csv("https://raw.githubusercontent.com/HigherEdData/Du-Bois-STEM/refs/heads/main/data/d_literacy_country.csv")

Recreating a Bar Graph

After successfully listing the data above, you should be able to see that it has data in two columns.

Each column is a variable:

  • country is a country name for 10 countries with Black people in the U.S. treated as a country.

  • illiteracy contains percent of people in each country who are illiterate.

Before creating a bar graph of the data, we need read the library ggplot2.

R 

library(ggplot2)

options(repr.plot.width=22/3, repr.plot.height=28/3)

source("https://raw.githubusercontent.com/HigherEdData/Du-Bois-STEM/refs/heads/main/theme_dubois.R")

Depending on our intention or circumstance, we adapt the size of a graph. The second line of code tells R that we want the width and height of the graph to have the same ratio that Du Bois used, 22 inches wide by 28 inches tall, with each divided by 3 so that it doesn’t display too big.

The third line tells R to use a specific Du Bois style. Examining Du Bois data portraits, we can see a style of format, color, and font text.

ggplot(d_literacy_country, aes(
  x = country,
  y = illiteracy)) +
  geom_col()

The code to create a simple bar graph is using the ggplot(df, aes(x=horizon_variable, y=vertical_variable)) + geom_col(). Within this expression, we set the parameters of which variable to be placed across the horizonal (using x=variable) and vertical axes (using y=variable). Typically, bar graphs have categories on the horizonal axis and the values as the vertical axis.

Challenge

Challenge 2: Horizonal Bar graph

Given that within the code above the variable is identify as x is horizonzatlly places and y is vertically placed. Knowing that, what can we expect if we exchange the placement of the code above?

ggplot(d_literacy_country, aes(
  x = _____,
  y = _____)) +
  geom_col() 
ggplot(d_literacy_country, aes(
    x = illiteracy,
    y = country +
    geom_col()

Now, our plot is starting to look more like the Du Bois original data creation. In the code from challenge 2, the categories are on x-axis (horizonal) and the rate/numbers are expressed on the y-axis (vertical axis). In the challenge above, we use the ggplot function to plot a horizonal bar graph of illiteracy rates across the observations (countries and Black Americans).

In the challenge above, we add a ggplot function followed by open parentheses to tell R that we will plot data from the d_literacy_country data frame with an “aesthetic mapping” aes specification that maps one column of data on the x axis and another column of data on the y axis.

After the close parentheses that tells ggplot we want to plot d_emancipation_dubois data with one variable on the x axis, and another on the y axis, we add a + and then a new line of code geom_col() that tells ggplot we want a bar graph based on summary statistics in the dataframe.

After looking at Du Bois’ version of the graph above, replace the _____ characters in the code cell below to plot the correct variable on the x-axis and the correct variable on the y-axis.

Sort the values & Adding color

In the bar graph you created above, can you tell what order the bars for each country are sorted by?

Du Bois sorts the bar for each country by its illiteracy rate from highest to lowest.

Du Bois also graphs illiteracy rate for Black Americans in a different color to make it easier to compare to other countries.

ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy))) +
    geom_col()

Compare the code above with challenge 2, what do you notice? The code above uses the reorder function to reorganize the country names on the vertical axis based on illiteracy.

ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Ireland"
  )) +
  geom_col()

Next, we have adapted the code to highlight a specific country: Ireland.

Challenge

Challenge 3: Reorder and specify bar color

Observing the PLate 02, what name is suppose to stand out? Update the below.

ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Ireland"
  )) +
  geom_col()
ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Negroes, U.S.A."
  )) +
  geom_col()

Adding Du Bois Theme

The style of our current figure does not quite match the original Plate 02. Can you identify 1-2 style characteristics missing?

Next, we will adapt the code to match the style, specifically: bar width, background color, legend, and other default ggplot elements ar different than those employed by Du Bois.

Bar width

We can easily edit the bar width by inputting numeric values within the geom_col(width = #) Try inputting .1 and .9.

Additionally, we can add theme_dubois() within geom_col(width = #) + theme_dubois()

Challenge

Challenge 4: Bar width and Du Bois theme

Add the changes we covered above.

ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Negroes, U.S.A."
)) +
  geom_col()
ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Negroes, U.S.A."
)) +
  geom_col(width=.5) +
  theme_dubois()

Changing bar color and font to match Du Bois theme

The function scale_fill_manual() allows users to adapt the colors of the fill based on values. Remember earlier, the legend printed colors based on TRUE/FALSE of country == “Negroes, U.S.A.”. We will on that previous knowledge to adapt the code: scale_fill_manual(values = c("TRUE"= "red", "FALSE" = "darkgreen"))

Changing the font to theme(text = element_text('serif'))

Challenge

Challenge 5: Changing colors and font style

Add the changes we covered above.

ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Negroes, U.S.A."
)) +
  geom_col(width=.5) +
  theme_dubois()
ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Negroes, U.S.A."
)) +
  geom_col(width=.5) +
  theme_dubois() +
  scale_fill_manual(values = c("TRUE"= "red", "FALSE" = "darkgreen"))+
  theme(text = element_text('serif'))

Labels

To add titles and subtitles to the graph, we use the labs function for ggplot2. We use the title and subtitle specifications with labs.

The title text needs to be enclosed in quotation marks. We use the code \n to tell R to put a “new line” break at different places in the title based on Du Bois’ titling.

Fill in the blank with your name in the title code below to show that the graph was recreated by you?

labs(
	title="Graph title"
	subtitle="2026"
)
Challenge

Challenge 6: Adding labels

Based on the Plate 02, update the title, subtitle, and your name.

ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Negroes, U.S.A."
)) +
  geom_col(width=.5) +
  theme_dubois() +
  scale_fill_manual(values = c("TRUE"= "red", "FALSE" = "darkgreen"))+
  theme(text = element_text('serif'))
ggplot(d_literacy_country, aes(
  x = illiteracy,
  y = reorder(country, illiteracy),
  fill = country == "Negroes, U.S.A."
)) +
  geom_col(width=.5) +
  theme_dubois() +
  scale_fill_manual(values = c("TRUE"= "red", "FALSE" = "darkgreen"))+
  theme(text = element_text('serif')) +
    labs(
        title = "\nIlliteracy of the American Negroes compared with that of other nations.\n",
        subtitle = "Proportion d' illettrés parmi les Nègres Americains comparée à celle des autres nations.\n\n
        Done by Atlanta University.\n\ngit ad
        Recreated by STUDENT NAME HERE\n\n"
    )