Predictive Data Madness

March Madness basketball

How Big Data and predictive analytics are changing the way we watch March Madness and do business

Data and sports is nothing new. Fantasy football, one of the most popular and data heavy activities, has been around since the late 1960s, and die-hard basketball fans have been using data to help them fill out their brackets years before the ease of the internet.

Now, imagine a basketball game: the players sprinting up and down the court, the squeak of sneakers on the hardwood, the swoosh of a clean basket, and the crowds’ raucous reaction.  It’s a simple sport at heart; all you need is a court, hoop, and ball. For a lot of casual fans, that’s as far as the game goes and when it comes to the insanely popular March Madness Bracket, those fans choose winners based on uniform color or the fact that Stanford’s dancing tree is a better mascot than…well, pretty much most of them. It would seem that predictive analytics has little to do with the sport. Well, yes and no.

For the serious fan, however, the game is much more complex. For those fans the court becomes a resplendent spreadsheet full of ever changing percentages, margins, and numbers. The tournament teams become columns of data; winning percentages, margins of victory, number of turnovers committed versus turnovers forced, offensive efficiency compared to defensive efficiency, field goal percentages, and so on and so on. Imagine the world as seen by Nero at the end of the Matrix, just all ones and zeros.

As the technology to handle huge amounts of data has developed, more and more people have been turning their attention to the wide range of applications for Big Data and sports. See, the power of Big Data analytics isn’t so much in what it can do with today’s data, but what it can do with the massive amounts of historic information. Kaggle, a collection of data scientists based out of San Francisco, have created a competition to see which predictive model is the best at determining not only the winners of each game along the way, but the percentage, or confidence of each pick.  While the predictions only use the data from the current season, the engines were built and perfected by looking back at two decades worth of stats. It’s only recently that the technology has existed that is capable of handling so much data in such a short amount of time.

Given the range of data that the Kaggle participants have, and the intense predicative engines they build, they are still unable to completely predict the perfect bracket. In fact, Warren Buffet was so sure of this fact the he laid down a billion of his own dollars betting against the perfect bracket to last year’s tournament.

So, what is it that we can learn from a data-filled March? Basically, it’s all about using the power of the past to predict the future. Knowing how a team has performed over the season, and knowing how the tournament has shaken out over the past two decades can go a long way towards seeing how a team will do and how the tournament will most likely play out. So while our crack team of data scientists at Nexidia might not make the Billion Dollar Bracket, we can certainly help companies look at a customer behavior over the course of several years and begin to see trends that will help focus everything from marketing strategies to entire business processes. By taking advantage of predictive analytics, everyone from sports fans to CEOs can begin to harness the history associated with their field and use it to make more informed decisions about future events.

Categories: Best Practices