Officer-Involved Shootings in Texas: 2016-2019

Texas law requires the Office of the Attorney General (OAG) to collect reports on all officer-involved shootings throughout Texas and then provide the summary of the individual reports in “The Peace Officer Involved Shooting Annual Reports.” The OAG digitally publishes these annual reports, which are extremely high-level summaries, with no findings or analysis of trends or intersectionality.

Texas Justice Initiative (TJI) reviewed the officer-involved shooting incidents data from law enforcement agencies in Texas between 2016 and 2019. This TJI publication improves upon the OAG’s annual reports by providing in-depth analyses with more granular information. Our report can be found on our website (see here). This report consists of two parts:

• The Data Summary section provides an overview of the incidents as intersectional information such as the total number of civilian severities per county and per race demographics.
• The Data Insight section provides TJI’s independent analysis of the officer-involved shootings, including notable patterns in the data.

This repository has the following information about the report:

• Jupyter notebooks for data preprocessing for the OIS datasets and the census datasets
• Jupyter notebooks for all the analyses in the report (Data Summary and Data Insight sections)
• Raw and preprocessed datasets
• Python script used for data preprocessing and visualization

Data

Raw

• Website: civilian and officer datsets downloaded from the TJI website (downloaded in June 2020).
• Census: census (Texas Demographics Center) and mortality (Texas Department of State Health Services) datasets

Interim

• OAG_report_summary.csv: Summary of OAG reports from 2016 to 2019
• census_county_race_2010.pkl: Preprocessed census data for populations by county (using the actual 2010 census data not the annual estimates)
• census_county_race_age_2010.pkl: Preprocessed census data for population by race, gender and age

Preprocessed

Preprocessed civilian and officer datasets both in the csv and pkl formats. See 1.0-hs-preprocess_OIS_data_OIS_report.ipynb for the details of data preprocessing process.

Notebooks

• 1.0-hs-preprocess_tx_census_data.ipynb: Prerocessing of the census data (c.f., the mortality data was scraped from the DSHS website).
• 1.0-hs-preprocess_OIS_data_OIS_report.ipynb: Preprocessing of the OIS data (both civilian and officer datasets)
• 1.0-hs-data_summary_OIS_report.ipynb: Analyses for the Data Summary section of the report
• 1.1-hs-data_insight_OIS_report.ipynb: Analyses for the Data Insight section of the report
• preprocess.py: Preprocessing script for all notebooks
• plot.py: Visulization script for all figures in the report

Figures

All image files are created as eps files. Figures_Notebook.zip has all figures created from the Jupyter notebooks (/Notebooks). Figures_Final.zip have the final version of the figures that are used in the report. These figures are identical to the notebook figures except for the colors in some.

Style Guide: Image styling guide that the graphic designer Ally Curtis created. The data insight and summary notebooks have most of the information in this guide.

Please contact Hongsup Shin for any questions.