Big Data

March Madness: 5 data sources that could predict the 2017 NCAA championship

The NCAA men's basketball tournament is one of the most exciting parts of the year for sports fans. Here are five data sources that could shed light on the final outcome.

bball.jpg
Image: iStockphoto/efks

For basketball fans in the US, the NCAA men's basketball tournament is a pivotal point in the college basketball season. Fanhood is on full display as March Madness descends on college towns across the country.

The 79th NCAA tournament will take place between March and early April 2017, with the championship game closing out the tournament on April 3 at the University of Phoenix Stadium. Back in 2016, it was estimated that some 28 million viewers watched the final game.

The NCAA tournament is very difficult to predict, as every team brings a renewed energy to the court. Upsets and Cinderella stories abound, with underdogs often taking out top teams early on. However, there are some statistics you can rely on to try and figure out who will come out on top. Here are five options for ranking NCAA tournament match ups.

SEE: Hiring kit: Data architect (Tech Pro Research)

1. FiveThirtyEight

Nate Silver's statistical analysis website FiveThirtyEight has been one of the most looked-to resources to predict everything from the outcome of political elections to the potential effects of snowstorms. The site's March Madness predictions show pre-game win predictions, and also has dynamic live predictions that change as a game is played. The site's own ranking system, Elo, combines with an excitement index and other team rankings to make its picks.

2. Georgia Tech LRMC results

Georgia Tech's H. Milton Stewart School of Industrial and Systems Engineering offers a host of predictions using a computer ranking system called LRMC that examines which two teams played, whose court they played on, and what the margin of victory was. Basketball fans can find a Bayesian model of LRMC predictions, along with more classic LRMC results and an LRMC option with no margin of victory.

3. SportsLine

SportsLine brings together data scientists, former bookmakers, and analysts to give their predictions on the tournament. The site provides standard statistical approaches alongside looks at major upsets and different possibilities.

Disclaimer: SportsLine is part of CBS Interactive, the parent company of TechRepublic.

4. Bing

Microsoft's Bing search engine has its own predictions for the tournament available as well. According to a Microsoft blog post, the company is using "intelligent machine-learning technology to analyze social and search signals, plus more than a decade of college hoops statistical data" to make its picks. For one thing, Bing is predicting a lot of early upsets.

5. Kaggle

As part of its yearly March Machine Learning Mania competition, data science community Kaggle is a great resource for tournament data. Community members are encouraged to post external data they are using in the predictive models, and often compare notes on what they are seeing.

Also see

About Conner Forrest

Conner Forrest is News Editor for TechRepublic. He covers enterprise technology and is interested in the convergence of tech and culture.

Editor's Picks