Election Analytics

View the Project on GitHub camilne9/election_analytics_blog

Gov1347 Election Analytics

Welcome to my 2020 election analytics blog! I will be updating my blog weekly exploring the 2020 US presidential election. You can reach out to me at via email (christophermilne@college.harvard.edu).


This is my first blog post for Harvard’s Gov1347: Election Analytics discussing swings in two party voting percentages in US presidential Elections. (9/13/2020)


In this post I consider how RDI growth in the third quarter of an election year can be used to predict the two party vote share of the incumbent party. I consider how the predictions based on RDI growth differ for incumbent presidents and non-incumbent candidates as well as how predictions differ for elections from 1960-1988 and from 1992-2016. (9/20/2020)


In this post I consider whether state level or national level polling is a better predictor for the US Presidential election by using state and national polling data for elections from 1972 to 2016. I then use state and national polling data to make a prediction for the 2020 presidential election. (9/27/2020)


In this post I consider the effect of incumbency on US Presidential election. Specifically, I look at how increases in federal spending by state between election cycles affect the two-party popular vote results in those states. (10/4/2020)

The Air Game

This post explores the effects of campaign spending on advertising on the US Presidential election. In particular, I use historical trends in how spending differences between parties in battleground states affects the result in those states and I use that to make a prediction about the 2020 election. (10/11/2020)

The Ground Game

In this post I consider the effects of voter turnout and voter demographics on Democratic performance in elections. I also revisit my model based on polls to update my prediction on the 2020 presidential election. (10/18/2020)

Shocks: Covid-19

In general we can consider how unpredictable or unprecedented “shocks” influence an election. For the 2020 election cycle, Covid-19 presents a massive example of such a shock. I explore how Covid-19 effects the 2020 Presidential election based on deaths per capita. (10/25/2020)

Final Election Prediction

This is my final post before the 2020 US Presidential Election. I will give my final prediction for the election including electoral predictions, state level popular vote predictions, and a national popular vote prediction. (11/1/2020)

Election Reflection

I reflect on the accuracy of my prediction and I consider the good and bad qualities of my forecasting model. I make hypotheses about what is responsible for my error. (11/23/2020)

Media Narrative

The media often makes claims about why the election result is what we observe. In this post, I examine the media claim that Hispanic voters were pivotal in 2020 and assess it’s plausibility. (12/10/2020)