Harshit Gaur

Predictive Winner - US Presidential Political Party

Introduction

Every four years, the US general election is held to choose the president and the vice president of the United States. The two major political parties competing in the general elections are the Democrats, and the Republicans. On the election day, the US citizen vote for their presidential choice in their respective electoral college in each county. The electors with majority in turn cast direct votes for the president.

Before the general elections are held, the presidential nominee from each party campaign across the country to gain support of the people. The campaigns are massive in the swing states of Florida, Ohio, and Pennsylvania, also known as the Battleground states. These campaigns cost millions of dollars, and there are designated campaign teams to plan and raise funds. The campaign teams from each political party leverage data analytics to understand and predict people’s stance on the political issues and candidates. The models and the insights help the political parties optimize their campaigns and target their outreach efforts.

In the actual US presidential elections, data from different sources like national databases, consumer preferences, social media, etc. are used to gauge the sentiment of the voters, target floating voters, and define strategies for advertisement. However, the scope of this project is to build regression model with Linear Regression, Stepwise Selection Regression, and Lasso Regression using historical election and demographic data to predict the winning political party in each county and understand the counties and swing states where the political parties should target their election campaign. The main questions that we tried to answer in this project are:

• What factors and attributes of a county influence the results of presidential election?
• Can we predict which party will win the 2016 presidential election in each county?

Analysis, Insights, Recommendations

Contact Me