MIDS Capstone Project Fall 2017

Personalizing Politics

Despite the consistent level of spending on midterm elections, voter turnout has always been significantly lower than the turnout experienced during presidential elections. Although much of this can be attributed to the magnitude of the election itself, we believe that a significant portion of this difference in voter turnout could possibly be attributed to a lack of visibility into the candidates themselves, along with a lack of interest from the voter to pursue this visibility. Throughout the entirity of the presidential elections, voters are fed very heavy viewpoints from the candidates, and these views will often drive voters to stick within their political affiliation. For midterm elections, this level of visibility into each candidates' political views is often far from the front page of the news. Seeking these views for each candidates can be a tedious and mundane task.

Our belief is that providing voters with an easily comprehensible understanding of which candidate best aligns with their personal political/social/economic views would help with driving larger number of informed voters to the polls.

To inform voters, we've developed a platform that intakes a user's political views for 20 political questions on various topics, and then return the most fitting candidates for the user's views. Along with each candidate's overall similarity to the user, we've also created a complex predictive analysis that allows us to present users with a relatively precise score regarding each individual candidate's likelihood of winning their election. This information would be used by a user to help understand how likely it is that their aligned candidate would actually win their state's election.

Although many of our analyses may seem fairly complex, we've ensured that we present this information to the user in an easy-to-understand format, using visualizations from tools like D3.js.

More Information

Last updated:

December 21, 2017