Machine
Football
2024 - PRESENT
AI-powered football analytics & fan engagement platform
My role: Digital Product Lead / UX / UI
The brief
Design a fan-facing web app that transforms professional-grade football data into something genuinely useful and engaging for serious supporters.
The platform needed to go beyond raw stats, delivering real insight, on-pitch context, and AI-generated content at scale, across 70+ leagues and thousands of players and clubs.
What makes it different
Data back on the pitch
Rather than presenting walls of metrics, I designed the experience around putting data where fans think about football; on the pitch itself.
The Threat Map shows not just where a player acts, but how often and how well.
The Game Model visual maps a team's full playing style across the 90 minutes, connecting directly to a radar graph that makes style legible at a glance.
Team Builder & player fit prediction
One of the most distinct features I shaped is the Team Builder, a tool that lets fans explore how players and playing styles fit together. It draws on cohesion modelling and style-matching algorithms to predict how any two players would combine, and surfaces the closest real-world team equivalent to any configuration you build.
AI-generated content at scale
We're training small language models on our data and infographics so they can genuinely understand matches, not just describe them. That feeds into an LLM layer that generates long and short-form content for every player, club, and match in the database, no matter how obscure the league or player. It's how we deliver depth and engagement at scale. I developed a robust templated system that allows written and graphical content to be combined in an easy to digest and well laid out format that can be automatically, or manually generated.
About the platform
Most football data platforms give fans numbers. Machine Football gives them understanding.
The product grew out of a B2B analytics engine built on six years of data covering every touch of the ball across more than 70 professional leagues. Rather than let that intelligence sit behind closed doors, we decided to build a fan engagement platform that could rival the likes of Sofascore and Transfermarkt, but with something those platforms can't offer: genuinely enriched, predictive, visually intelligent analysis.
My Role
My role was to define what the product should do, how users would move through it, and how it should look and feel. I led on feature definition, user flows, and all design, working closely with the data science and engineering teams who built the underlying models and trained the AI.
The design approach
The visual language had to work hard. A dark-first interface keeps the on-pitch visuals sharp and legible. The radar graphs, threat maps and cohesion webs needed to be immediately readable without instruction so I focused on colour, weight and hierarchy to do the heavy lifting. On mobile, the experience is card-based and progressively disclosed, letting users drill into as much or as little detail as they want without ever feeling overwhelmed.
The homepage pulls everything together: live match context, weekly editorial picks, AI-surfaced predictions, and ranked club and player lists all in a single scrollable feed that rewards the curious fan.
Launching April 2025
Currently in closed beta. Logo by external collaborator; all product strategy, UX, UI, colour palette, typography, and visual language by me.
By the numbers
70+
Professional leagues covered
6 yrs
Of historical data
Every touch
Of every ball, scored and ranked