Hillary or Trump? Democratic or Republican? Who will win? It is hard to see into the future, but we will try to predict the results of the 2016 US election. We collected a lot of data from open databases and created a dataset. It is available for everyone, so you can download and analyze it. We are sharing it with you because it is always hard to find the source of the studies, analysis, or charts used in articles. We hope that we can help you a bit and make your life easier. If you find interesting facts or correlations, please let us know. Share the science and share the data!
You can download the dataset here: US president and election data from 1916
Hillary Clinton Was Born on a Weekend, and Donald Trump Was Born on a Weekday
Why is this information important? Have you ever thought about what the day type is of the candidate’s birth date?
Weekend vs. Weekday Birthdays
Above you can see the distribution of day type for birth dates. One quarter of presidential candidates’ birthdays are on a weekend, which is different from the expected 2/7. However, what about the winners? Does it matter which type of day were they born on? Yes, it does.
Win Percentage of President Candidates Born on Weekdays (Monday-Friday)
Win Percentage of Presidential Candidates Born on Weekends (Saturday and Sunday)
Bad News for Hillary
As you can see, according to our dataset, if the candidate was born on a weekday, there is a 47.8% chance of winning. If the candidate was born on a weekend there is only a 20% of winning. So candidates who were born on a weekday have a higher chance of winning the election.
Let’s A/B test it: Weekdays vs. weekends
According to the A/B test, we can say the following:
“Born on Weekdays seems to be better than Born on Weekends by 139%. Perhaps (89% chance) Born on Weekdays is somewhat better but you can't be sure.”
We know we cannot predict the results of the election with this data, but it is interesting to analyze the data and discover hidden relations and correlation
How Else Can We Predict the Winner?
Check the Data of the States
In which state should the candidate get the majority of the votes to win the whole election? From the dataset, we built a decision tree. Right now, we cannot share the pictures of our tree, but if you are interested in the full application, sign up for the newsletter, and we will inform you when it will be published. It is coming soon!
The chance of winning is 40.98%. That means that in 25 rows, there is a winner there have been 25 presidents since 1916. Now keep digging. As we calculated, if the candidate gets more than 47.85% in New Mexico, he or she would have 88.89% chance of winning the election. Moreover, if the candidate gets more than 47.85% in New Mexico and more than 50.3% (win) in Nevada, there is a 100% that he or she will win the election. This calculation is based only on historical data, but who knows!