Murders By Country
Visualization Critique
On February 21, 2020, security.org posted a page of various different data visualizations that summarized murder in the United States. Among these different visualizations there was one that specifically compared the U.S. to other countries worldwide to compare murder rates. This visualization is pictured below and can be found at: https://www.security.org/resources/homicide-statistics-by-state/
The Data
The data for this visualization was taken from a dataset formed by The World Bank, an organization that works towards sustainable solutions to reduce poverty in developing countries. The database focuses on intentional homicide and was drawn from the UN Office on Drugs and Crime’s International Homicide Statistics Database.
The data represents murders per 100,000 people in 17 different “high-income” countries with recent data.
Analyzing the Visualization
According to Kieran Healy, problems with visualizations come in three different varieties: aesthetic, substantive, or perceptual.
The visualization that security.org created does not have many aesthetic issues. The design of the visualization is pleasing to the eye and has consistent design choices. The one thing that does take away from the visualization aesthetically are the tangle of lines that connect the country to the data. Edward R. Tufte argued that the viewer should get the greatest number of ideas, in the shortest time, with the least ink, in the smallest space. These lines do not achieve this. Instead they waste ink and it takes the user time to trace the lines from the data to the country or vice versa. The focus should be more on the numbers in this case rather than the map of the countries, which is not relevant to the data that is being presented.
There are a couple of substantive issues with this visualization. The creators of this visualization stated that the countries represented were “high-income” countries. However, it is not immediately clear what this means. The metric used to determine what is considered a “high-income” country is not mentioned. If one uses median household income as a metric, perhaps Mexico, Brazil, Russia, and Indonesia should not be included. They all have lower median household incomes compared to the rest of the countries presented. Additionally, Mexico, Brazil, and Russia all have significantly higher murder rates. The inclusion of these countries can lead to biased opinions. If one were to remove these three countries, U.S. would have the highest murder rate by a large margin.
Perceptually, this visualization does well in some aspects and not so well in others. The use of colors in the visualization is well thought out. Color schemes should be chosen carefully, and in this case they used unordered hues to separate the different continents. If they were to use a different gradient, the viewer may be left wondering what different colors mean. In this case, the colors do not show any sequential ordering which works nicely in this context.
The shortfalls of the visualization perceptually, is evident in the layout of the numbers. The data points are scattered above the map of the world and display the numbers at different heights. This can be confusing because the height has no indication of value. It is also hard to find the number for a specific country or specifically, find the highest or lowest number. These types of things should be immediately evident.
There are a few other noteworthy criticisms to be made about this particular visualization. While data-to-ink ratios are not entirely indicative of a good visualization, it can often be a tool to use when assessing the efficiency of a visualization. In some cases it is okay to sacrifice the ratio in order to make a visualization more memorable. Sometimes a memorable piece with a lot of “chartjunk” can get a message across in an efficient way. This is true in the case of the “Monstrous Costs” visualization shown below which is quite memorable indeed. I don’t believe that this is the case for this visualization. The actual map of the world is an example of this. The map does not add any value. The goal of the creators was to compare the U.S. murder rates to other countries and the map does not aid the viewer in achieving this outcome. Lastly, nothing pops. No conclusion really jumps out at the viewer and instead they are left to wonder what the point of the visualization actually is.
Redesign
Based on the criticisms that I made about the visualization created by security.org, I took the numbers that they presented and redesigned the way that they were shown. The redesigned graphic can be viewed below:
The most evident change in this visualization, is that it is a bar chart. The way that the numbers were presented previously was not sufficient to compare among countries, which was the ultimate goal. A bar graph allows the murder rates to be much more efficiently compared.
Another change that I made was to exclude Mexico, Russia, and Brazil from the data. I don’t believe that these countries, were in the same category as the countries that security.org was trying to show. They were trying to compare the U.S. to other countries that are comparable in income. These three countries have significantly lower median household incomes than the U.S.
One thing that I kept from the old visualization was the color scheme. The unordered hues allow the viewer to separate the countries into their respective continents by use of non-sequential colors.
Lastly, the goal of my visualization was to compare the U.S. to similar countries in terms of income. The comparison immediately pops. It is clear that the U.S. has much higher murder rates than its comparable peers.
