Midterm Report Peer Review #9
Description
This project aims to predict which candidate will win a swing state in the 2020 U.S. Presidential election. They plan to analyze data providing demographic, socioeconomic, and electoral data to reach their conclusion. The relevancy of this project is important for future elections for candidates to improve their chances of winning a key swing state.
One quality of their project I like is their data visualization techniques for all of their data. Four extremely detailed correlation matrices we’re generated to show correlation between party alignements and age, gender, occupation, education level, and many more features. This is a key component for their analysis since these trends will identify which groups of people a candidate needs to invest for time on. Their approach to prevent under/over fitting using 5-fold cross validation is smart since we tend to generalize or stereotype someone’s political affiliation too easily; this bias must be avoided when developing a model for this kind of problem. Finally, their plans for future work seem very well though out and they plan to incorporate data from the results of the 2020 election.
I think including plots of the results of the models would be more effective than explaining them in words. It would be hard for the reader to understand your results without visualizing them as well as it was done for the preliminary exploratory analysis. The name and explanation of the method used was not provided, so it is hard to tell what future models they will use to try to improve their results. The reader has no idea what model is used and why R^2 and MSE are relevant to your analysis. I also think the report needs more explanation instead of just referring to other documents or files for the reader to do on their own.