I work at the intersection of data science and human behaviour, helping institutions benchmark performance, evaluate innovation ecosystems, and make evidence-based decisions.
I'm an analyst at EFIS Centre, a Brussels-based not-for-profit organisation specialising in research and innovation policy and governance for EU-level clients, including the European Commission. My day-to-day involves data analysis, quantitative modelling, report writing, and helping EU institutions understand how national and regional governments perform across innovation, resilience, sustainability, and competitiveness dimensions.
My academic path - spanning mathematics, psychology, and data science - gave me a somewhat unusual toolkit: quantitative rigour, a feel for behavioural nuance, and the technical skills to bridge the two. I'm most energised by work that sits in that gap between raw data and the human decisions it's meant to inform.
Outside of work, I follow Formula 1 and basketball (and frankly any other sport) with possibly excessive analytical attention. To relax I enjoy working on modelling kits and doing origami.
A selection of analytical projects I've contributed to, spanning scoreboard development, qualitative research, data visualisation, and sports analytics.
Contributed to the design and production of the ESS for the European Commission, developing R scripts for automated visualisations, generating country reports, and writing analytical commentary on EU27 member state performance.
Contributed to the design and production of the EIS for the European Commission, developing R scripts for automated analytical commentary production and generating country reports.
Conducted methodological analysis of indicator selection rationale and data quality for the Eco-Innovation Index (EII) revision, including hands-on validation of datasets.
Led the design and execution of a large-scale, AI-based qualitative analysis pipeline for the ST4TE research project, processing ~400 interview narratives across multiple research questions using multi-step synthesis workflows and prompt engineering.
ST4TE aims to provide a comprehensive view of the drivers of the twin transition (TT), the inequalities that emerge or are widened by it, and a set of policies to build greener, more equal and more productive societies
Built an expected points framework for the NBA from scratch, training an XGBoost classifier on 4M+ shots spanning 1997–2020. The model predicts shot success probability from spatial, contextual, and player-level features, enabling game and season simulations.
Assessed Double Poisson, Bivariate Poisson, and Dixon & Coles models, and applied them to the 2017–18 Premier League season, estimating team-level attack and defence parameters and simulating the full league 10,000 times to assess predictive accuracy.
The technical and analytical capabilities I bring to my work - spanning programming, data analysis, and policy research.
I also enjoy writing articles from time to time. Here's a list of selected pieces I've worked on.
Adapting football's xG revolution for the NBA - a shot-level model trained on 4M+ attempts that separates skill from luck and reveals which players truly create value.
Sports AnalyticsTesting three Poisson-based models on the 2017–18 EPL season — simulating the league 10,000 times to find out whether simple maths can capture the beautiful game's chaos.
Sports AnalyticsA home for small interactive tools and visualisations I'm building. Both pieces below are in progress — check back soon.
An interactive map of the EU — pick a stat, see how member states perform comparatively.
In progressHover over any spot on the court to see the model's expected points per shot, built on 4M+ historical attempts.
In progress