As weeks after their official ending, the US elections are still making headlines every day, it is our aim to have a closer look at the results from the US elections, and examine underlying trends in voting behaviour. The data can be found under the following link:
https://www.kaggle.com/etsc9287/2020-general-election-polls This data is supplemented with other datasets, such as poll data and voter turnout data, depending on the question at hand.
More specifically, we will attempt to answer the following questions:
- What was the general outcome of the election, and how are votes devided?
This question is analysed by visualising the spread of votes using maps, and analysing areas which are dominated by Trump or Biden voters respectively. In particular, the swing states which are crucial states which determine the outcome of the election, are highlighted in the discussion.
- How did the turnout rate change between 2016 and 2020?
The 2020 elections boasted the highest turnout rate of past century. We establish that people were particularly keen to vote in several key states, and examine what the outcome was on the overall election results.
- How do factors like race, unemployment, salary and gdp influence votes?
After establishing the correlation between the factors and the share of a county's votes that go to a certain candidate, a few particularly interesting datapoints are examined more closely through scatterplots.
Our methodology is primarily EDA-based. This means that we attempt to tell our story through visualisations and maps. Certain statistical methods, such as correlation calculations, are still used. However, due to the large topic to be covered, we made an active decision to not employ machine learning techniques for prediction, as we felt that the limited resources available could be used to convey a more powerfull message.