Using Data to Predict Soccer Outcomes
DECEMBER 14, 2016 • BLOG POST
Using machine learning API platform, Haven OnDemand, HPE analyzed soccer match outcomes
As any sports fan knows, there are really two levels to any competition. There is the game on the field, being played in real-time with unpredictable factors, such as weather conditions or injuries. And then there is the game managed with data and statistics that are regularly used by coaches and players, as well as by fans, broadcasters and others with a vested interest in the outcome.
Sky Sports, the largest sports broadcasting group in Europe, has collected a huge amount of data for the sporting events it covers. HPE has had the capability and the functionality to do predictive analytics for a few years. Utilizing the data available from the partnership between HPE and Sky Sports, would it be possible to predict the outcome of an English Premier League soccer match?
Using its machine learning API platform, Haven OnDemand, HPE was able to analyze the statistics of the two teams and draw conclusions about the possible outcome based on that data. Information about players and the teams’ past performance was analyzed by Haven OnDemand, and the machine learning platform was able to successfully predict Manchester City’s 2-1 win over Manchester United.
While it is all in good fun to use this technology to anticipate the results of a sports game, the ability to predict outcomes, such as the likelihood of winning a sales opportunity or even to assess upcoming events using location-based analytic forecasting, is of benefit to all industries. However, every data set is different, and data analysts may not know which prediction model is best for their specific data set.
For the sports industry in particular, predictive analytics is more than just forecasting score outcomes. Sports are a huge business and athletes demand large salaries. The data collected for individual players can help with analyzing the trajectory of future abilities and provide feedback for owners on whether or not the player is worth the financial risk. For coaches, whose jobs depend on team performance, predictive analytics can suggest which players to use in certain game situations. Furthermore, clubs can even use predictive analytics for game day sales forecasts of branded merchandise and food products based on team rosters and other conditions.
However, when it comes to the collection of data, the sports industry isn’t different from any other industry. Almost all industries have more data than they know what to do with, yet, they aren’t always adequately equipped to extract actionable insights. The difficulty is finding the right tools and the people with the necessary skills to make data work for each individual business. Until recently, success has been limited to only those companies that have been able to hire the best data scientists and that can provide access to very specialized technology. What HPE developed with Haven OnDemand is the ability for everyday developers to easily create models from training data to identify the right prediction technique to use, and to generate predictions and recommendations from these models while at the same time providing developers with the control over which model they ultimately wish to use. Developers simply supply historical data where the outcomes are already known to train the platform to find the most likely predictions.
Predictive analytics is proving to be a game changer. With technology like Haven OnDemand, HPE customers have on-demand access to powerful machine learning predictive analytics APIs. To get a prediction using your data, simply upload it to HavenOnDemand.com, specify the questions to be answered, and get predictions or recommendations in minutes. The platform is automating all the complex data science and doing all the heavy lifting to make data work for the customer.