This project was completed solo. The goal of this project is, by using data obtained by the US Census Bureau, to build a model that will predict the margin between the percentage of Democratic and Republican votes, and to identify its corresponding most important features. Ideally, this model should be useful for anyone working in politics to provide insight as to where to focus resources. Multiple supervised machine learning algorithms (LASSO, Ridge, Random Forest, Gradient Boosting, and Support Vector Machine regression) were conducted, with the Gradient Boosting-derived model to have the best overall performance. However, the error rate is still high enough that the utility of this model may be minimal, and the model should be further developed.
This repository contains:
- Code - Python code written in Jupyter notebooks
- Source Data - Downloaded from the US Census Bureau website
- FinalPaper_DSC540_StevenJordan.pdf - Project Report
- FinalPresentationt_DSC540_StevenJordan.pdf - Project Presentation File
A video presentation of the report can be viewed here: https://www.youtube.com/watch?v=3YTtVXiw0yk