Precinct-level Election Results

Description: A map of 2016 general election results with precinct-level data, combined with demographic data about race and average income taken from the 2014 American Community Survey, built with Leaflet. The vote results are collected by scraping the Secretary of State elections website, and can be automatically merged with the map data as needed

See the data_cleaning dir for the scripts used to get the vote data, combine it with demographic data, and calculate summary statistics. The get_data dir also contains the script used with QGIS to calculate demographic data for voting precincts based on overlapping census tracts.

Published: Nov. 8 2016

Lives Here: http://www.myajc.com/atlanta-neighborhood-2016-presidential-election-results-map/

Hosted Here: http://ajcnewsapps.s3-website-us-east-1.amazonaws.com/2016/precinct-election-map/

Installation

Run the following commands from the top level directory of the app.

Install node dependencies npm install

Install nvm and use the right version of Node npm install -g nvm nvm use

Install Python dependencies (if you don't have virtualenv it's a good idea to install it) pip install -r requirements.txt

Setup

To run the scraper and regenerate the map with data from a Clarity elections site:

1. Edit the constants at the top of clarity_live.py to fit the parameters of the race you're interested in scraping. The scraper should work for race data provided by Clarity, a vendor used by more than 20 states to show live precinct-level results.
2. You may need to install PhantomJS separately and pass the path to the binary in the clarity_live.py script. Get PhantomJS
3. Run the clarity_live.py script to grab the most recent results and regenerate the map: npm run update_data. You must have a recent version of Firefox installed. Make sure to check that Selenium is compatible with your Firefox installation. Pauses have been added to avoid Selenium crashing so the script can take a little while to run. The script is far from perfect so it may throw errors.

Run the development server npm run serve-dev

Deployment

To deploy the map with live updates on election night, simply start a cronjob that runs python data_cleaning/2016/clarity_live.py and npm run deploy-production in succession. Use npm run deploy-staging to deploy to the staging environment.

If you run into trouble with the shell process spawned by cron not being able to access the virtualenv, follow the instructions in this StackOverflow answer.

Notes

• Frontend code lives in the assets directory
• The python scripts used to collect data and update the map live in the data_cleaning directory

Data

• Census data about income and race from the ACS 2014 survey
• 2014 precinct shapefiles from the reapportionment office
• 2014 census tract shapefiles