Predicting Presidential Candidate Sentiment

Usage

• Run Driver.py to obtain formatted CSV
• Run prediction.py to make predictions

Introduction

Twitter’s early promise as a political tool has become ingrained as a political reality... The amount of discursive access to politicians is unprecedented in the past century of American politics. (Vann R. Newkirk II, "The American Idea in 140 Characters", 2016)

It's common knowledge that Twitter is a popular choice for those who wish to voice their opinions on a variety of subjects -- especially politics. Therefore, if we mine tweets for the sentiment surrounding important elections, can we predict what candidate will win the election? Or, how many votes each candidate will receive?

Objective

The goal of this project is to predict U.S. presidential candidate sentiment on a state-by-state basis by tweet data.

Contributors

This is a research project developed by Nicole Dillon, Marie Dolleman, Alexis Kulash, Kelsey Olson, and Jennifer Steffens in conjuction with Drake University professors Eric Manley and Tim Urness.