eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
There are a few ways to anticipate the result of a sporting event: 1) by emotion, which is pulling for a favorite team to be victorious, whether or not it’s actually the better team; 2) by educated guess, which is using one’s own knowledge and intuitiveness to project a winner; and c) by data science, which entails utilizing big data analytics to determine who should rightly be the event champion.
All these viewing methods, of course, are commonly found in places like Las Vegas, yet they still amount to mere guesses in the end.
However, with all the sophisticated new IT now available in the big data analytics sector, one would think that more accurate sports prognostications could and should be made available. After all, huge business problems—such as projecting where oil wells should be drilled and predicting sales spikes based on time of year and product sales history—are being solved every day by these spiffy systems.
Quest for Science-Aided Clairvoyance
SAP, the Germany-based business software maker that has found renewal with its in-memory HANA database and the new analytics apps that run on it, decided to take a crack at predicting the victor in the Feb. 2 matchup between Super Bowl XLVIII foes Denver Broncos and Seattle Seahawks. In its quest for clairvoyance, SAP not only analyzed football season statistics but also factored in natural language opinions and analysis from social networks like Twitter, LinkedIn and Instagram.
Using data from NFL.com for the entire postseason, SAP crunched the numbers—from passing yards to defense to special teams—with the company’s analytics solutions in order to predict the outcome for this climactic championship tiff. Data visualizations were created with SAP Lumira to illustrate the findings and how SAP determined a winner. You can check out a video and images here.
“We wanted to have some fun with this,” Nic Smith, SAP’s senior director of marketing for Analytics, told eWEEK. “We used the products of a couple of recent acquisitions—Lumira for data visualization and KXEN InfiniteInsight for crunching the numbers—to come up with conclusions that might make the difference in the game.
“So there are some advanced analytics and cool algorithms going on there. Then we have a broader portfolio of what we call ‘predictive analytics,’ in which our customers can add in their own data points.”
Added Factor: What the Social Networks Are Saying
The last piece, Smith said, was the addition of what the fans are saying about the game. SAP’s Netbase software, which understands the sentiments and inferences within natural languages, was used for that part of the job.
“We are adding in streaming data and social media by looking into what people are saying leading up to the game,” Smith said. “This includes the sentiment of the conversation, who is saying what about which team, and how strongly people feel about which team will win.”
SAP Uses Own Big Data Analytics to Project Super Bowl Winner
As of the afternoon of Jan. 31, Smith said, Seattle had the edge when it came to fans tweeting and blogging about the team and how much better they think the Seahawks are than the Broncos. But pregame chatter does not a final score make.
Smith said that special-teams data can turn out to be the most important when it comes to determining the final result. Denver clearly came out ahead there. For example, how many times has a field goal, onside kick or muffed punt figured prominently in the final outcome of a game? The analytics say “more often than you might think.”
Key Takeaways From the Analysis
Key takeaways from the analysis, according to Smith:
–The Seahawks have a slight advantage on rushing yards, but the Broncos’ passing yards blow the Seahawks away, giving the Broncos the advantage when looking at total yards.
–Analyzing opponents’ scores, the Seahawks have the advantage with the best defense in the league, keeping opponents to much lower scores than the Broncos throughout the regular season.
–A head-to-head look at value-added by special teams throughout the season gives the Broncos the advantage.
–Using predictive algorithms, the most weighted characteristic of a winning team, showed special teams as the most contributory factor.
–Based on the variables of rushing yards, passing yards, total yards, defense and special teams, SAP’s Predictive Analysis gave the winning edge to the Denver Broncos.
When all the available data was compiled, the social networks were scanned and the data illustrations were completed, Smith said, SAP came up with a projected final score for the game: Denver 26, Seattle 23.
Lessons Can Be Learned by This Use Case
If this actually turns out to be the final score, then we might be seeing Smith as a special guest Monday morning on the network talk shows.
Even though this is all in great fun, there are some good lessons here about the value of analytics for many use cases.
Despite all of the above analytics, the qualifier “might” is always going to be the key factor in any sports use case. After all, in football it often takes only one key fumble, interception, tip-toe pass catch or missed tackle to decide who wins and who loses. The analytics all go out the window when something like that happens.
Still, it’s interesting to see how a case for a winner (or loser) is built by factoring in as many data bits as possible from the two teams before they butt helmets on the field Sunday afternoon in front of tens of millions of television and Web viewers.