Chicago-based sports analytics company STATS has developed a new platform for football clubs to analyse game video and performance data, using computer vision and AI to ease the discovery issues plaguing analysts in an age of data overload.
STATS Edge is an application which allows coaching and scouting staff to search through video clips by asking natural language questions, thanks to a combination of computer vision (so the platform can understand what it is seeing) and machine learning-driven search. These efforts have been driven by a quickly growing data science team, led by director of data science and ex-Disney (owners of ESPN) researcher Patrick Lucey.
STATS CEO Kenneth Fuchs explained the move at the time of release, saying: “It’s increasingly difficult for clubs to process and analyse all the relevant data points available and distil complex information into timely, accurate insight.
"We’ve applied computer vision and artificial intelligence to bring new levels of context and efficiency to teams already using video for their match preparations. It’ll save them time, allow them to develop proprietary results, and help get them the information to make, share and track the best decisions.”
The idea is to give analysts an easier route to insight, as they can now search through STATS' huge trove of data and video from across Europe's top leagues and the Champions League. STATS now tracks around six million data points per professional game.
While building Edge, STATS worked with at least three Premier League teams to develop the platform through design thinking workshops, intended to identify and address their biggest pain points.
The result is an analytics platform which allows analysts and scouts to quickly find the relevant content, annotate it and share insights with players and colleagues in the form of a 'playlist' of video clips.
Speaking to Techworld, Richard Henderson, chief revenue officer at STATS explained that the idea is for Edge to "provide a common language for different functions across user personas in coaching, the academy and the board to communicate more effectively".
So for example, Edge allows a user to retrieve all of a team’s counter attacks which led to goals, or all of the shots taken by a striker. As well as being useful for game planning, Edge can effectively be used by scouts to assess transfer targets in more detail.
Henderson says that pre-Edge the industry standard is to analyse an opponent's previous five games, as they scramble to create a game plan in a short window.
The problem with this is it can be a limited sample size. Especially if a key player has been injured for the previous five games, but is due back to play your team.
Importantly, Edge allows for deeper analysis, so not just looking at all the through balls a player made during a game, but also what the result of that pass was. STATS is also building out its 'what if' scenario modelling capability within the platform.
The entry level price for the Edge platform is $50,000 (£38,000) a year for multiple logins.
Find your next job with techworld jobs