Police Data Accountability

This project has two components and serves as a proof of concept for ECCSC – a Chicago non-profit offering first responders as an alternative to the police.

Most violence intervention groups primarily utilize Chicago Police data; however, ECCSC would like to use Citizen data to evaluate the accuracy of the Chicago Police Data, and consider using both data sources to compare their own first responder activity against. Their primary concern is that the Chicago Police data may not always categorize incident type correctly.

Project Structure

1. Record Link:

Project EZ runs an analysis on two primary sources used to track crime incident data for Chicago. We apply a record linking algorithm to match crime data extracted from the Chicago Data Portal’s API to reported incidences of crime pulled from crowdsourced application Citizen. We match incidences based on geo coordinates, date, and time. This enables ECCSC to evaluate Citizen reports of crime compared with Chicago’s Police response to crime and assess any severity differences in reporting.

2. First Response Activity:

Pulls down data from a google survey which ECCSC uses to log first response activity. Cleans data and generates latitude and longitude coordinates to map into a map of Chicago using Dash. ECCSC can visually assess their first response activity for internal operations and to demonstrate their value proposition to prospective funders.

Modularized into Four Components

1. Chicago Police Data

• Chicago data API pull
• Visualization of ECCSC data using Dash
• Command line interaction to select which component of project to run

2. Web Scraper for Citizen Data

• Built automated daily citizen pull (on local machine)
• Storage system for preserving csv files aggregating by month of crime incident and key search terms

3. ECCSC Crime Prevention Activity

• Pull down google sheet ECCSC data
• Find latitude and longitudes based on inputted survey data
• Record link algorithm

4. Transform and Analysis

• SQL database connection for each data source
• Clean incoming data and data types
• Record link algorithm

Interact with EZ Project

To run our project from command line: cd ~/proj-ez sh install.sh (if local machine requires different command line arguments for virtual environment, then enter following commands to run program) cd ez bash project-requirements

Expected output:

Interaction:

• Asks for user input, if they would like to run record link analysis, map data, or exit program
• Once user inputs, echos user choice and runs corresponding command
• If still in script after command is run, reminds user of possible inputs From Link Records Analysis:
• Prints the start and end dates captured by Chicago Crime data and Citizen. Prints number of days captured between the two datasets which overlap and can generate a match
• Writes all matches to a csv file called “match_file.csv” From ECCSC Map visualization:
• Displays map of first response activity from ECCSC data generated from google forms survey. Click cursor on marker to display detail of crime type.