This repository was created to display our submission to the Data Science for Good: Center for Policing Equity Kaggle competition.
Developers: Joseph Bentivegna, Ariana Freitag, Matthew Grattan
Advisor: Professor Sam Keene
Institution: The Cooper Union for the Advancement of Science and Art
A link to the Kaggle kernel can be found here.
The approach we took to analyzing the provided data is focused primarily around the challenge stated in the problem statement: automating the combination of police data, census-level data, and other socioeconomic factors. Since the ultimate goal of the CPE is to inform police agencies where they can make improvements by identifying areas of racial disparity, our kernel aims to provide a tool to the CPE that combines police data and census-level data into a comprehensive dataset that is ready for out-of-the-box unsupervised machine learning. We believe that applying unsupervised machine learning algorithms to police and census data can prove useful to identify areas of racial disparity because they can capture inter-district discrepancies in quality of service.