Data and replication code for analyzing crime on street segments with and without building removals in Baltimore, MD 2014-2019

This data publication includes data from a study in Baltimore, MD to examine how removing vacant buildings affects violent and property crimes. This study includes data from in 2014-2019. Housing data were obtained from the Baltimore City Department of Housing and Community Development, data from reported crimes were obtained from the Baltimore City Police Department via OpenData Baltimore, and demographics were obtained from the American Community Survey (ACS) - U.S. Census Bureau. Data include 775 block faces (which is the area between two street intersections, including the space around the street between buildings facing each other) that had building removals (treatment) as well as 524 block faces that did not have any building removals (control). This package includes the location of the treated block face segments as well as the centroids of all treated and control block faces as both shapefiles and geopackages. Controls were randomly assigned to treated block faces. After removing spatially near candidates, there were more treated units than controls, and we randomly assigned 251 controls a second time to treated block faces; all treated units had a matched pair control and associated treated-control identifier. Also included are the annual tabular data for each block face which include data such as average crime rate, number of deconstructions/demolitions, whether or not a violence initiative was put in place for that area, information regarding possible transformation zones, police district details, and demographic data. In order to see if the results were sensitive to the treatment/control matching, we applied a second random matching procedure to re-match treatment and control block faces. All treated units had a matched pair control and associated treated-control identifier called "subclass," hence the two tabular data files contain the same data with the exception of the subclass variable which provides that random match for both matching procedures. The R code used to run these analyses is also provided as an *.Rmd file.
The purpose of this study was to evaluate whether or not vacant building removals (either demolitions and/or deconstructions) in Baltimore, MD had a demonstrable reduction on various types of crime via a quasi-experimental difference-in-differences (DID) analysis.
For more information about this study and these data, see Locke et al. (2023). These data were published on 06/22/2023. Metadata updated to include reference to newly published article on 08/02/2023.