FBI Crime Data Explorer Scraper

This is a simple Python script that scrapes the U.S. arrest data by state and by agency using the Federal Bureau of Investigation's Crime Data Explorer (CDE) (API). I originally wrote this script for work to benchmark FBI's Uniform Crime Reporting (UCR) data against the data we have acquired at the Criminal Justice Administrative Records System (CJARS) at the University of Michigan (for current data holdings, see here). I'm assuming there might be similar codes out there but here is another one in case some one is looking for U.S. arrest data by offense type. So please use responsibly! 😉

Output

The run.py file will save 3 different types of .xlsx files (~100 files altogether):

• ucr_ori_crosswalk.xlsx: Crosswalk of Agency ORI
  • API Endpoint: 'sapi/api/agencies'
• arrest_by_agency_*.xlsx: Agency-level arrest data for each sate by offense type
  • API Endpoint: 'sapi/api/data/arrest/agencies/offense/{ori}/all/{min_yr}/{MAX_YEAR}'
• arrest_by_state_*.xlsx: State-level arrest data by offense type
  • API Endpoint: 'sapi/api/data/arrest/states/offense/{state}/all/{min_yr}/{MAX_YEAR}'

Install

First, clone the repository:

$ git clone https://github.com/jaycatsby/ucr_scraper.git

Make sure you have all of the required packages (in virtualenv preferably):

$ pip install -r requirements.txt

Run

Register

If you haven'd done so already, sign up for an API Key: https://api.data.gov/signup/

Edit settings.py

• Set API_KEY in line 3 to what you received in the registration email (e.g.): API_KEY = 'AGKQGIJPQEOJH!LNHPIJh31-9ujpfkn-h9h'

• (Optional) Set RAW_PATH: By default, all of the data will be saved as .xlsx files in raw folder of the current directory.

• (Optional) Set MIN_YEAR: By default, starts from 1985. I initially set this to 1975 to see if there would be differences in coverage but from my initial glance, most of the data seem to start in 1985.

• (Optional) Set MAX_YEAR: Currently data up to 2018 is available. Edit as see fit.

• (Optional) Set MAX_WORKERS: Please be responsible! By default, set to use 2 processes

Scrape

After editing settings.py, run run.py

$ python run.py

Features

• Stata Support: After scraping, run clean_arrest.do file to generate *.dta files of the arrest files in ./raw