Football data analyst — Turin, IT
Data visualisation is what I do best. At Math&Sport I work with event and aggregated data — and, more recently, tracking — producing visual and editorial output distributed to DAZN Italy, Cronache di Spogliatoio and Radio Serie A. Independently I point the same craft at recruitment: transfer corridor models, role clustering, custom metrics, published with every method open at The Cutback.
Work with me
So I take one defined question — a position to fill, a market to choose, a target to stress-test — and run it through a documented pipeline until it becomes something a sporting director can act on. Fixed scope, fixed price, methodology handed over with the result.
Everything I'd build for you, I've already built in public. You can judge the work before you pay for any of it.
Selected work — The Cutback
An atomic VAEP-based pipeline across six seasons of event data, measuring transfer success rates corridor by corridor — which origin leagues make adaptation easiest, and which don't.
A complete mock recruitment process: team similarity to shortlist target leagues, corridor risk, heatmaps and passing-skill profiling — from Stiller to Camavinga, Fernandes and Avdullahu.
Eighteen verdicts I published last summer, checked against what actually happened — the wins, the honest misses, and the ones that still need time.
A scouting brief built the way a club would commission it: define the profile, filter candidates across leagues, and land on a recommendation backed by metrics.
Technical work
Each notebook and metric below maps onto a stage of the recruitment workflow. Together they're the pipeline I run for clients.
Stage 1 — Define the role
Grouping players into data-defined roles — the position feature behind any profiling system.
Stage 2 — Pick the market
Rating stylistic closeness between clubs to identify the leagues where a target is most likely to translate.
Stage 3 — Evaluate players
A metric to rate defending from event data — methodology, validation, application.
Estimating composure under pressure without pressure-event data.
An acceleration estimate from event data, validated against Gradient Sport's physical tracking output.
Two competing proxies for a quality traditional stats miss — scouting dribblers without tracking data.
About
I read the game through data, but my first training was in people: an MA in Cultural Anthropology from the University of Turin still shapes how I interpret collective behaviour on the pitch — and how I write about it for the analysts, scouts and editors who read The Cutback.
At Math&Sport I work with aggregated and event data daily in a production environment, turning it into visual and editorial output on deadline.
Fluent
Event data — six seasons of metric building, profiling and analysis on top of it
Working
Tracking data — recent production work, plus the limits of what's publicly available
Stack
Python · pandas · scikit-learn · Streamlit · VAEP & xG modelling · data viz
Languages
Italian (native) · English (professional)