Political Party Affiliation Predictor

When we decide to vote during elections, the outcome is relatively clear. One ballot, one vote towards the candidate chosen. However political involvement can take many different forms beyond voting such as monetary donations, volunteering, campaigning, etc. The impact on politics by individual contributors is more complicated to measure. During the 2014 election cycle, only 0.23% of the United States population gave more than $200 to political candidates, committees, or parties. However, two-thirds of the money that went towards senatorial candidates and half of the money that went towards congressional candidates, came from the 0.23% of the population that contributed. With the raising in cost of election campaign, every single dollar counts.
Which political party are you more likely to donate to?


About the Political Party Affiliation Predictor App

Companion Blog

Find all of our indepth analysis and insights on what we mined out of this data. Understand the top 1% contributors vs the bottom 99%.

The Data

Under the Federal Election Campaign ("FEC") Act, all contributions made to federal candidates, committees, or parties over $200, must be reported. These records are made publicly available through the FEC and several other websites. The data can be found here and spans 2011-2012 and 2013-2014 election cycles.

The Model

The model was built using Azure Machine Learning Studio and R. The data was processed, cleaned, and prototyped in R. The final model was trained and hosted on Azure Machine Learning Studio and features a multi-class logistic regression.

The Website

The predictive model is wrapped around the Django framework and JavaScript. Django is a Python based web framework used for rapid deployment of websites.