This project is by no means an endorsement or an indication of my political beliefs

Thesis)

Based on party incumbency and cumulative inflation over the previous 4 years, I am predicting Donald Trump to win the election with 58% of the electoral college.

Due to the restrictions of my data and model, a 95% confidence interval places Trump’s electoral college win between 38% and 78%. This is a wide margin, but it skews over 50%

Exploratory Data Analysis)

After assembling and cleaning the data, I created this plot to visualize trends based on whether the candidate was Democrat or Republican, whether their party is incumbent, and most importantly- cumulative inflation over the previous 4 year period.

This graph demonstrates that Republicans do better both as incumbents and non-incumbents when inflation during the previous period is high.

Creating a Refined Model)

I created a linear regression model to predict the current election by focusing on data points for Republican, non-incumbent candidates

The result of my model was a positive correlation with a p-value of 0.04 for y-intercept, and 0.03 for total inflation percent. The model has an R^2 of .60 which I found to be satisfactory

The plot above shows the data points and regression line for Republican candidates when their party is non-incumbent. A vertical line is shown at 21.2% cumulative inflation to represent the current circumstances. This line intersects with the regression line at 58% which is what the model predicts Trump’s electoral college winning percent to be.

You may have noticed that the model is influenced by a high outlier in the upper right quadrant of the plot. This outlier is from the 1980 election in which Ronald Reagan won 91% of the electoral college with cumulative inflation of 45% during the previous administration. Removing this outlier results in the following plot:

The above plot and associated regression model result in a predicted electoral college win of 59% for DJT.

Lastly, this plot shows Trump’s predicted electoral college % regardless of incumbency and without removing the 1980 election results. It results in a 60% predicted electoral college win, but an R^2 value of just 0.28

Conclusion)

Based on the model I have created, it is likely that Trump will win the 2024 election with about 58% of the electoral college or 312 electoral votes. My model has a wide confidence interval, but I find the correlation between inflation, party incumbency, and a candidate’s electoral college winning percent to be convincing.

Discussion)

Donald Trump ended up winning 312 electoral votes to Kamala Harris’ 226 votes in the 2024 presidential election. This puts Trump at an electoral college win of 58%, which falls exactly in line with the model I created and the outcome I predicted. It is quite neat that my prediction was exactly the same as the actual outcome, so I am happy with the prediction model I created. The only thing I wish I could have done better was narrow my condifence interval.

Resources)

I gathered cumulative inflation data using the following website: https://smartasset.com/investing/inflation-calculator

I also used Chat GPT to occasionaly fix coding bugs and errors, but avoided using it in creating entirely new code