Original


*Source: https://www.reddit.com/r/dataisbeautiful/comments/cz58sb/blood_type_distribution_in_the_united_states_oc/*


Objective

The Visualisation provides an overview about the percentage of the population of United States with different blood groups. Looking at the Pie Chart provides us an estimate of which blood groups are most common and which ones are the rarest to find at the time of blood donation.

The visualisation chosen had the following three main issues:

  • Pies and Doughnuts: The audience cannot compare the numbers drawn from the visualisation. Viewer has to compare the angles and would be curious to know the figures behind the various sectors of the chart which will keep the audience curious about knowing it. Group A and O are approximately the same size and there appears to be a minor difference between them which may lead to bad communication with the audience.

  • Colour Associations and Trifecta Check-up: The purpose of using the specific colour cannot be figured out. It is just for the purpose of differentiating the different sections. Although diffent saturation levels are used to communicate the positive and the negative of the named group but no clue over the use of a specific colour assists the decision. Moreover, there is no hint over what purpose is solved by seeing the visualisation, why the different groups have been plotted, what we monitor from such visualisation and it triggers questions regarding the same which needs to be enlightened.

  • Gestalt Laws - Figure Ground Principle: The variables are over-emphasised than the actual data. The audience is more attracted towards what is written in the front than concentrate over the data.

Reference

Code

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

library(ggplot2)
library(dplyr)
blood <- data.frame(Groups = c("A-", "A+", "AB-", "AB+", "B-", "B+", "O-", "O+"),
                      Perc = c(6.3, 35.7, 0.6, 3.4, 1.5, 8.5, 6.6, 37.4))

p1 <- ggplot(data = blood, aes(x = Groups, y = Perc))
p1 <- p1 + geom_bar(stat = "identity", colour = "#000000", fill = "#FFAAAA") +  geom_text(aes(label = paste(Perc,"%",sep="")),  nudge_y = 1) + labs(title = "Blood type distribution % of 326,481,533 Population in US", subtitle = "AB- is the rare blood group which is 0.6% of the total population of the United States", caption = "Source: https://www.wikipedia.org")

Data Reference

Reconstruction

The following plot fixes the main issues in the original.