Thinkful Student Uses Data Science to Predict Touchdowns

I bet you 5 bucks they score a touchdown on this play. The last time I heard someone make a wager like that was during Super Bowl LII. Not only would I guess incorrectly and leave my friend's Super Bowl party five dollars poorer, my beloved Patriots would end up losing to the Philadelphia Eagles. All in all - not a great day.

Amongst fans of the National Football League, these type of spontaneous bets are incredibly common. According to one estimate, over $90 billion is wagered each year. And while most us will use our intuition in these situations, new powerful data tools can take out most of the guesswork. In fact, Albert Troszczynski, a student in Thinkful's data science bootcamp has proven that it's possible to predict whether or not a team will score a touchdown with incredible accuracy. His model which he developed for his Thinkful capstone project is correct 96% of the time.

To build his model, Albert used a dataset he found on Kaggle. The data set contained information about every regular season play for every team between 2009 and 2017. In total, he analyzed 407,688 plays and more than 100 different features of those plays.

Albert's model which uses logistic regression is not incredibly accurate - it's also incredibly fast. It's able to make a prediction in less than half a second.

For Albert, the Thinkful capstone was a great way to combine his passion for football with his newly developed data science skill set. To read more about his project including which 5 features (i.e yardage, time, down) best predict touchdowns, I encourage you to check out the Thinkful blog.

This post was sponsored by Thinkful.

Want to learn more about Thinkful? Read what alumni have to say on SwitchUp.