Designing a recommendation engine for Football
Data plays an ever important role in the success of football clubs, from event data in game analysis to searching hundreds of thousands of players to support player recruitment. ScoutLab aims to augment this process further, helping the decision making of football scouts with machine intelligent data insights that remove bias and present a shortlist of prospects simply and quickly.
I head up design at Frontier and worked closely with Ben Blackmore (Frontier’s Tech Lead) to frame the project around “How might we help football scouts use data to shortlist players quicker?” and develop a Minimal Viable Product within a short timeframe.
Data-driven insights for Football Scouts
The concept of ScoutLab took inspiration from e-commerce recommendations as found in Amazon and media giants Netflix. Both companies excel at using data to personalise the customer experience, with serendipitous recommendations that require little user input. With ScoutLab, we wanted to make the activity of searching thousands of prospects more intuitive, this was a core focus of the system design.
Traditional scouting data-tools make this search quite problematic, in that the search query is based on assumptions of what data qualities represent a certain type of player. For people, it’s much easier to make reference to another player who they feel embodies the right qualities to fit the recruitment brief. If a club want’s to sign a ball-wining midfielder, they could search ‘Kante’ or if they seek a wing-back with strong distribution they could start with Tottenhams ’Trippier’. This ‘anchor’ player can play for any club, the important part is ensuring they embody certain characteristics. These qualities are what represent the players profile and are what the system bases its recommendations on.
ScoutLab is responsive in nature, affording scouts in the field a way to quickly shortlist and compare players from across footballs top tier divisions. ScoutLab is the first spin-off ‘Lab’ from ‘Frontier Labs’ and takes cues the parent identity, branching out the colour palette while making Libre Baskerville the primary typeface.
Frontier Labs hope to build upon ScoutLab, with further experiments in Sport and other industries.