❮ Projects page PennPraxis

Spatial Analysis Project:

The United States is undergoing a death-by-gun crisis: in recent years, there have been roughly 15,000 gun homicides and 20,000 gun suicides annually. Lawyers, data scientists, economists, and urban designers have been tackling this issue both qualitatively and quantitatively. Scant spatial analysis has been conducted to date with regard to measuring the strength of gun laws, however, especially within public policy and legal academic circles. This project aims to shift the current paradigm and understanding of gun policy by using spatial analysis to present new, deeper understandings of how our gun laws work and to spark an interdisciplinary conversation among scholars, data scientists, and policymakers. From mapping gun laws in all fifty states to applying machine learning analyses to large urban areas, the project seeks to identify those specific gun laws—within the context of their geographical locations—that have the greatest impact in reducing both homicide and suicide rates.

Data available:

Our main dependent variables document fatal and non-fatal shootings. Currently, these datasets span 2009–2018. The sources for this data include:

1. Federal Bureau of Investigation (FBI) Uniform Crime Reports,
2. FBI Supplementary Homicide Reports,
3. Center for Disease Control WISQARS,
4. Open Data sources from thirty-three US cities, and
5. Gun Violence Archive (GVA).

Independent variables include various state gun laws. The sources for this data include:

1. Brady Center for Gun Violence Brady Scores,
2. Boston University School of Public Health Gun Law Database, and
3. a self-developed gun law database.

Most of our dependent variables have been geolocated. For the FBI data, we matched all reported incidents to the reporting agency. For Open Data and GVA data, we have geolocated the data to the street level and to the coordinate level.

Maps and Reports that will be created:

We expect novel insights on the effectiveness of gun laws in reducing homicides and suicides either on the local, state, or federal level. We hope to identify areas where gun laws make a difference. For example, we hope that machine learning techniques can be applied to investigate the violence-reduction impact of laws such as the Gun-Free Schools Act, which prohibits loaded or unsecured firearms within a school zone. We also expect to distinguish regions with similar characteristics, such that predictions can be more precisely tuned to particular parts of the country. Such insights would help us better understand how gun laws function in different settings. Lastly, we hope to visualize the implementation or annulment of gun laws across time and space and to detect correlations with gun violence.

How the maps and reports will be used:

The maps and reports will be used for three purposes. First, we plan to share our results with both academic and public forums through publication and debate. Second, we hope to use them as foundations for further analyses in assessing the strengths of gun laws. Because gun laws vary geographically (e.g. each state has its own set of gun laws), it has been challenging for researchers to measure the outcomes of each of these laws effectively. Third, we hope to present our findings to legislators at the local, state, and federal level. For example, we may be able to testify at legislative public hearings that center on gun legislation. Our ultimate goal is to use the spatial representation of the death rates by guns and their relationship to geographically-based gun laws to prompt changes in those laws throughout the country.