A set of quick-and-dirty scripts used build a list of automotive dealerships in the USA.

git clone https://github.com/simon-wenmouth/dealerships.git

cd dealerships

virtualenv venv

source venv/bin/activate

pip install -r requirements.txt

On a per manufacturer basis run the download scripts, i.e.

cd dealerships/manufacturers/lincoln/

./download_data.py

less ../../../data/lincoln.json

This repository contains the data these manufacturer's dealerships.

NADA reports that there were 16K (or so) new car dealerships and the total of the above table is greater than 20K which means that there are dealerships with multiple franchises (i.e. they are in more than one file).

As such, I have merged the various manufacturers into a single normalized file (based on URL) which has identified 19,171 different dealerships. Evidently a better heuristic is required.

• 2015-05-29 Initial revision pushed to GitHub.
• 2015-05-31 Created normalized JSON files.
• 2015-06-01 Created normalized CSV files.
• 2015-06-03 Added CSV files containing
  • DNS record data

  - Web sites meta-data for various dealer sites I was able to identify through the DNS system.

• 2015-06-05 Initial revision of dealership list that are hosted by Cobalt Nitra (4721 sites).
• 2015-06-06
  • Initial revision of dealership list that are hosted by Clickmotive (2471 sites).
  • Initial revision of dealership list that are hosted by Dealer.com (9731 sites).
  • Initial revision of dealership list that are hosted by DealerFire.com (199 of 351 sites).