Geospatial Computation Final Project

A spatiotemporal analysis of mass shootings in the United States.

Introduction

Gun violence is generally characterized as a mass shooting when there are four or more casualties. According to a CNN article, Monday's supermarket shooting in Colorado marks the seventh mass shooting in the US in seven days (Berlinger, 2021). It's hard to tell how this compares to past weeks since the US government does not centralized database of gun violence data. Analysis of where, when, and how these shootings are occurring is essential as pressure on lawmakers mounts to enact more firearm legislation. Since laws, values, and norms vary by state and region, this data is inherently spatiotemporal and should be analyzed as such.

Data and Methods

Data

There is no open-source centralized federal data base of gun violence. Many independent information sources have put together their own data bases such as the Gun Violence Archive (GVA), the Washington Post, Vox, and Mother Jones. For this project, two lists of mass shooting data were pulled from the GVA and Mother Jones. The database from the GVA includes 2000 mass shooting incidents over the last 5 years. The database from Mother Jones is not as extensive, but includes 121 of the most widely reported mass shooting incidents from 1982 to 2021. Both datasets were geocoded and entered into ArcGIS Pro as point layers.

Most Widely Reported Mass Shootings from 1982-2021. Click on points for more information. Source: Mother Jones

More comprehensive list of mass shootings from 2016 to 2021. Source: GVA

Methods

    Spatiotemporal analysis was carried out on the GVA data to investigate trends over the last 5 years.

    • Initial assessment was done using a data clock, which visualizes numbers of occurrences by month over a target time period.
    • Hot Spot Analysis was done to examine clustering in the data.
    • Emerging Hot Spot Analysis was carried out to examine trends in the clustering of data over time.
    • A spacetime cube was created that contains bins of data by spatial and temporal neighborhood.
    • This spacetime cube was visualized in 2D to examine clusters over time.

An emerging hot spot analysis was also carried out on the Mother Jones dataset. Analysis using a space time cube was less useful as there were many fewer data points. However, comparing the two time periods is useful for examining how recent trends compare to trends over the last 40 years.

Results

Data Clock

Data clock of mass shootings from 2016 - 2021

The data clock shows that the amount of mass shooting incidents seem to increase over the five year time period. The highest volume of mass shootings in a month occurred in June 2020, with 95 incidents. There don't seem to be any clear trends by month.

Hot Spot Analysis

Map displaying hot and cold spots of mass shootings from 2016-2021

The hot spot analysis revealed clustering trends of mass shootings across the US over the last five years. Hot spots were clustered in Southern California, Texas, and New Mexico while cold spots were clustered near east coast and Midwest cities like New York City, Philadelphia, and Chicago.

Visualizing the Space Time Cube

Map displaying emerging hot and cold spots of mass shootings from 2016-2021

The emerging hot spot analysis revealed large hot spots forming around eastern Texas and the Atlanta area, among other locations. A few cold spots were forming, most notably in southern Texas.

Discussion

Some factors that were suspected to contribute to the probability of a mass shooting were population density, strictness of gun laws, and guns per capita. These factors were compared to incident hotspots to see if data followed this predicted trend.

Conclusions

There are many factors that contribute to mass shootings in the US that couldn't be considered in this project. Taking action to prevent future mass shootings requires a good understanding of why these incidents occur. Continued analysis of temporal and spatial trends is an important part of action against these tragedies.

It is important to acknowledge that this analysis was limited by the lack of comprehensive data on mass shootings in the US. While these results are important, drawing actionable conclusions requires a complete data set.

Works Cited

Berkowitz, B. & Alcantara, C. (2021, March 22). The terrible numbers that grow with each mass shooting. Wall Street Journal. https://www.washingtonpost.com/graphics/2018/national/mass-shootings-in-america/

Berlinger, J. (2021, March 23). The Colorado attack is the 7th mass shooting in 7 days in the US. CNN. https://www.cnn.com/2021/03/23/us/7-mass-shootings-7-days-trnd/index.html

Follman, M., Gavin, A., & Pan, D. (2021, March 22). US Mass Shootings, 1982-2021: Data from Mother Jones’ Investigation. Mother Jones

Gun Violence Archive. Mass Shootings. Retrieved from https://www.gunviolencearchive.org/mass-shooting

Morris, K. (2020, March 04). Here are the most (and least) gun friendly states. Retrieved March 28, 2021, from https://www.zippia.com/advice/least-gun-friendly-states/

Data clock of mass shootings from 2016 - 2021