Football data is becoming increasingly important, and clubs are thus hiring more and more data specialists to study it. Some clubs have even started to use Moneyball schemes – for example Brentford, which has recently been promoted to the English Premier League and maximizes its reliance on data.
But how can a football fan benefit from football data while on the Watson online news platform?
Many people are enjoying Fantasy Football nowadays, a game that lets you create your own team and compete against your friends and colleagues. What better way is there to compete than with a data-driven decision-making aid that lets you pick your players?
Football fans always know best, but with this tool they can finally back their convictions with data. The challenge here is to develop a simple analytics app that even non-techies can use and that draws on openly available data.
Datahouse, a data science services provider, came up with this challenge for the Watson online news platform. Datahouse will monitor the project from a data science perspective and deliver the finished app to Watson.
Football fans always think they know best when talking about their splendid game. To help them along, we designed an app to manage the data themselves and back their convictions with objective facts rather than subjective opinions. Our app allows fans to do their own analyses and gauge the true value of a player's performance in the top five European football leagues (France, Germany, Spain, England, Italy).
The app aims to attract more football followers to Watson’s website and thus increase the number of visits. The more website visits it generates the more attractive Watson will be to advertisers.
The app offers users three ways to analyze a player's performance:
- Choose from seven position-based performance indicators, whereby the app then finds the best-matching player.
- Predict the most valuable player for each position based on the most relevant performance indicators for the position.
- Visualize the two selected indicators so that users can analyze the player's performance.
Potential next steps
- Repeat the task at the end of the current season to obtain a larger sample size
- Expand the data to include more leagues in addition to the five that have been selected.
This challenge used data from https://fbref.com/ about the top five European football leagues. Part of the challenge was to select the available data and to categorize, organize and store it in a database.