St. Louis, Missouri consistently tops the charts as one of the most dangerous places in the nation. While this is not untrue, St. Louis’ status as an independent city skews the crime statistics as the city and county are separate entities. In order to dig deeper and contextualize the statistics, this dashboard was designed to examine and map the crimes committed in St. Louis County in the first three-quarters of 2020. By analyzing this data, a more complete picture can be gleaned of the crime rates and types committed in the overall metropolitan zone, as well which areas are more plagued by crime.
Process & Materials
The first step to creating this dashboard was to acquire a data source. For this project, I downloaded a CSV from GIS Open Data Portal of 2020 crime data from each precinct within St. Louis County. Crimes reported in this file are homicide/non-negligent manslaughter, rape, robbery, aggravated assault, burglary, larceny, and motor vehicle theft. The data required very little manipulation, so I imported it into Tableau Public to create a visualization of the trends and used a feature there to clean it. After reviewing the data I grouped the “Crime” column into broader categories, for example, Larceny $200 and over, Larceny $50 to $199, and Larceny under $50 were combined into “Larceny.” This allowed the data to be studied at a higher level.
I was really inspired by Gabriel Dance and Tom Meagher’s article “Crime in Context,” published in 2016 on the Marshall Project’s website. The study focuses on providing more information on politicians’ various claims that crime rates are either rising or falling, asking “Violent crime is up in some places, but is it really a trend?” I appreciated how they made the data publicly available so that anyone with pertinent skills could analyze it for themselves. One of the charts that stood out displays a graph of violent crime rates since 1975, and as framed by Obama and Trump. I also reviewed data previously published by St. Louis County organized into a “St. Louis County and Municipal Crime Map,” which views crimes by the municipality on a graph and map.
Explore the visualization: St. Louis County Crime Statistics, 2020
I began creating my own visualization by working with graphs of the data, first looking at a breakdown of the types of offenses committed. A pie chart was used to show which type of crime, i.e. homicide, larceny was occurring most frequently. I then moved to a bar graph showing the number of crimes committed within each city in St. Louis County, organized from highest to lowest. The lower right-hand of the dashboard has a stacked area graph showing the number of crimes committed from the beginning of the year to September 1 when the data ends. Types of crime are color-coded within the chart to show viewers the frequency of all crimes and each offense specifically.
This data set included latitude and longitude values for most complaints logged, omitting some for victims’ safety and privacy. I used Tableau’s ability to plot these points on a map to examine where hotspots are located, color-coded by offense type. An added filter allows users the option to view one type of crime plotted at a time.
The color scheme I chose to employ in this visualization is a gradation from yellow to red, with yellow being the least severe crime, orange representing more severe crimes, and red showing the most severe and violent offenses. Originally, this color scheme was modeled after a traffic light (shown in Figure 2), with green being more minor crimes. However, when viewed on the map, a peer review noted that they felt the green color meant safer areas of the city when in fact the safest areas would have no plot points at all.
A prototype of this visualization shows a much different layout, with the original color scheme. The “Offenses per Jurisdiction” section was organized alphabetically rather than by descending crime rates. The graph over time simply showed the total number of crimes throughout the year but provided no data on the breakdown of type.
Reflection & Future Direction
Overall, this dashboard does a good job of presenting the crime activity in the County of St. Louis. The main questions that this visualization answers are where are crimes being committed, and what type of offenses are occurring? In a future visualization, I would desire demographic data on who is committing these crimes and who the victims are.
Ideas that I had that would contextualize this data even further are comparing the crime rates on a map to the poverty levels, average household incomes, unemployment rates, average education levels, and the number of police officers in each precinct. St. Louis is one of the most segregated cities in the nation and this is evidenced by its racial, economic, and public safety divide. Adding these filters to a future study of crime rates provides background to the viewer on how the disparity in the County results in violent and desperate outcomes.