Click the Original, Code and Reconstruction tabs to read about the issues and how they were fixed.

Original


Source: HowMuch.net, a financial literacy website (2019).


Objective

In One Chart: State Tax Revenue by Source was created by Juan Carlos and published on 11 November 2019 by HowMuch.net. The objective of the visualisation is to display the different types and amounts of tax revenue the US Government received from all 50 states in 2018. The values represented in the visualisation are cumulative revenue totals for 25 different taxes.

The target audience for this visualisation is individuals or corporations interested in the impact of the wealth tax plan proposed by Democratic presidential candidate Elizabeth Warren.

Using the visualisation, Carlos (2019) has attempted to represent that the most significant source comes from the overrepresented revenue of income taxes. However, the remainder of the top five revenue sources is not present in the plan.

Carlos (2019) argues that adding additional tax to wealthy individuals may not provide funding for the proposed programs and that there are other potential sources of revenue the wealth tax plan does not acknowledge and could help pay down the federal debt.

The visualisation chosen had the following three main issues:

  • The visualisation displays too many variables. High numbers of variables make it difficult for the reader to determine relationships between revenue sources.
  • The use of area and size in the circle has a low accuracy for representing numerical data. Comparing the different size and areas of variables make it difficult for the reader to determine any relationship between sources.
  • Inconsistent use of symbols and visual indicators. Not all areas of the visualisation contain consistent imagery. The use of symbols, images, or visual indicator distracts the reader from any relationships the visualisation is attempting to represent.

References

Code

The following code was used to fix the issues identified in the original.

# Load libraries
library(ggplot2)
library(dplyr)
library(ragg)
library(rsconnect)

# Import data set
onechart <- read.csv("in_one_chart.csv")

# Check structure of data frame
str(onechart)
## 'data.frame':    25 obs. of  4 variables:
##  $ tax    : chr  "Property Taxes" "General Sales and Gross Receipts Taxes" "Alcoholic Beverages Sales Tax" "Amusements Sales Tax" ...
##  $ amount : int  20079471 317422936 6795036 8151881 22457399 48262883 125587 12549600 19502424 47124945 ...
##  $ percent: num  0.0195 0.3079 0.0066 0.0079 0.0218 ...
##  $ type   : chr  "Property taxes" "General Sales Taxes" "Selective Sales Taxes" "Selective Sales Taxes" ...
# Convert type to factor
onechart$type <- as.factor(onechart$type)

# Confirm factor conversion
str(onechart)
## 'data.frame':    25 obs. of  4 variables:
##  $ tax    : chr  "Property Taxes" "General Sales and Gross Receipts Taxes" "Alcoholic Beverages Sales Tax" "Amusements Sales Tax" ...
##  $ amount : int  20079471 317422936 6795036 8151881 22457399 48262883 125587 12549600 19502424 47124945 ...
##  $ percent: num  0.0195 0.3079 0.0066 0.0079 0.0218 ...
##  $ type   : Factor w/ 6 levels "General Sales Taxes",..: 5 1 6 6 6 6 6 6 6 6 ...
# Convert data into 2 decimal format
onechart$percent <- onechart$percent*100 %>% round(digits = 2)

# Generate horizontal bar chart
p1 <- ggplot(data = onechart, mapping = aes(x=reorder(tax, percent), y=percent)) # Load data into plot
p1 <- p1 + geom_bar(stat = "identity", aes(fill = type), # Select bar chart geom aesthetic
                    position = position_nudge(y = -1.5)) + # Adjust plot y axis position  
  coord_flip() + # flip the x axis to display horizontal
  labs(title = "A Breakdown of all States' Taxes Revenue", # Add title label
       subtitle = "Taxes Revenue was $1.031 Trillion in 2018", # Add subtitle label
       y = "Percent (%) of Taxes Revenue by Category", # Add axis label
       fill = "Tax Type") # Add legend label
p1 <- p1 +  geom_text(aes(label=paste0(percent, "%")), size = 3) + # Add data labels
  theme(plot.title = element_text(size = 20, face = "bold", hjust=-4.75)) + # Make plot title 20pt and bold
  theme(plot.subtitle = element_text(size = 16, face = "bold", hjust=-1.68)) + # Make plot subtitle 16pt and bold
  theme(axis.ticks.x=element_blank(), axis.text.x=element_blank(), # Remove axis ticks and text
        axis.title.x = element_text(face = "bold")) + # Make x axis label bold
  theme(axis.title.y=element_blank(), # Remove y axis title text
        axis.text.y = element_text(face = "bold")) + # Make y axis labels bold
  theme(panel.grid.major = element_blank(), # Remove plot grid lines
        panel.grid.minor = element_blank(), # Remove plot grid lines
        panel.border = element_blank(), # Remove plot border lines
        panel.background = element_blank()) # Remove plot background
p1 <- p1 + theme(legend.position = c(.975, .5), # Adjust legend position
                 legend.justification = "right", # Right alight legend content
                 legend.text = element_text(size = 7), # Adjust legend text size
                 legend.title = element_text(size = 9)) # Adjust legend title size
p1 <- p1 + scale_fill_manual(values = c("#88CCEE", "#999933", "#661100", "#117733", "#332288", "#AA4499", 
                             "#44AA99")) # Add colour-blind palette

Data Reference

Reconstruction

The following plot fixes the main issues in the original.