Finding insights from past election data.

The outcome of the 2020 US Presidential election is becoming less and less predictable by the day. Will a vaccine be available by November? How many people will (be able to) vote? How will wildfires, riots, and the coronavirus change people’s voting behavior? There’s not even agreement on how many swing states there are — 6, 10, 11, perhaps 12 or more?

There are many opinionated arguments to be made, but there’s hardly a rigorous way to analyze how unprecedented current events will impact voting habits. We can, however, analyze past elections to measure the impact of various attributes on presidential elections and find insights into how 2020 may turn out.

Data

We’ll use this dataset shared on Kaggle by a data scientist at Nvidia.

Each row of the data represents a US county — with 3,143 counties — and includes how many votes candidates got in previous elections, as well as over 100 other attributes, including figures on race, education, earnings, poverty, population, age, health, weather, and more.

#data-science #election-2020 #trump #artificial-intelligence #machine-learning