Voter Sentiment Analysis Based on Tweet History

Location

Room 2904 A

Session Format

Paper Presentation

Research Area Topic:

Computer Science - Large Data Computing (Big Data)

Abstract

Under the leadership of Dr. Hayden Wimmer and Dr. Jeffrey Kaleta, this undergraduate senior Information Technology capstone work aims to research the correlation of election contests with outcomes based on a majority vote and social media applications. Twitter data related to the 2016 presidential election is extracted via Twitter’s Application Program Interface (API) using an automated script written in the R programming language and an automated script written in the Python programming language. Search terms to query the API include candidate names, campaign slogans (official and unofficial), and hashtags for caucuses and primaries. The data is stored in JSON format on a cloud based storage service to be parsed for analysis. The JSON data will be pulled into industry leading business analytics and data science software in order to 1) forecasting to predict the outcome of the caucuses and primaries, 2) analysis of events such as candidate dropouts or news headlines to see how the occurrence affects voter sentiment, and 3) association market basket analysis of the mention of candidates names by verified candidate Twitter accounts to see how the general public’s opinion is affected by the interactions of candidates on Twitter. The analysis will take into consideration the volume of tweets per topic of interest and the attitude of the tweets to determine positive or negative weight on the candidates/events. We anticipate that social media can be a predictor of election contests based on conversation quantity, content, and event situations.

Presentation Type and Release Option

Presentation (Open Access)

Start Date

4-16-2016 4:00 PM

End Date

4-16-2016 5:00 PM

This document is currently not available here.

Share

COinS
 
Apr 16th, 4:00 PM Apr 16th, 5:00 PM

Voter Sentiment Analysis Based on Tweet History

Room 2904 A

Under the leadership of Dr. Hayden Wimmer and Dr. Jeffrey Kaleta, this undergraduate senior Information Technology capstone work aims to research the correlation of election contests with outcomes based on a majority vote and social media applications. Twitter data related to the 2016 presidential election is extracted via Twitter’s Application Program Interface (API) using an automated script written in the R programming language and an automated script written in the Python programming language. Search terms to query the API include candidate names, campaign slogans (official and unofficial), and hashtags for caucuses and primaries. The data is stored in JSON format on a cloud based storage service to be parsed for analysis. The JSON data will be pulled into industry leading business analytics and data science software in order to 1) forecasting to predict the outcome of the caucuses and primaries, 2) analysis of events such as candidate dropouts or news headlines to see how the occurrence affects voter sentiment, and 3) association market basket analysis of the mention of candidates names by verified candidate Twitter accounts to see how the general public’s opinion is affected by the interactions of candidates on Twitter. The analysis will take into consideration the volume of tweets per topic of interest and the attitude of the tweets to determine positive or negative weight on the candidates/events. We anticipate that social media can be a predictor of election contests based on conversation quantity, content, and event situations.