Data analyst based in Bologna, working on data warehousing, ETL pipelines and data quality at UNI, the Italian standardisation body. Former Blue Book trainee at the European Commission (2024–2025).
I am interested in public-sector and official statistics: turning administrative and open data into reliable, documented indicators.
Tools: SQL · Python (pandas, scikit-learn) · R (sdcMicro, sdcTable) · dbt · DuckDB · PostgreSQL · Power BI · data validation and data quality checks · Git and CI
| Project | What it shows |
|---|---|
| eu-stats-pipeline | Eurostat SDMX ingestion with file-level lineage, dbt models on DuckDB, 27 validation rules including regional–national coherence, an ESS-structured quality report, a generated data dictionary and a NUTS 2 dashboard |
| marche-regional-stats | A Power BI report on the Marche region: an eight-page regional profile with a comparison switch, a map indicator selector and a province drill-through; business-register checks on constant boundaries, NUTS 3 indicators from Eurostat, SARIMA forecast evaluated out of sample (repository) |
| eu-silc-sdc | Statistical disclosure control in R (sdcMicro, sdcTable) on EU-SILC public microdata: re-identification risk, three protection scenarios, weights that do not reveal the region, release checks that stop the run if a file leaks, information loss by source on poverty and income indicators, cell suppression audited with p% intervals from linear programming, published report |
| ai-act-standards-map | Which high-risk requirements of the EU AI Act (Articles 8–15) are covered by harmonised standards, and when: a curated dataset with a public source on every row, links that rest on judgement marked as such, a weekly check that opens a review issue when a CEN-CENELEC page changes, and a one-page site (repository) |
| juve-momentum-index | Football data platform (FastAPI, PostgreSQL, Next.js): scheduled ingestion, Elo ratings and a Momentum Index on a dashboard, with the limits of its backtest documented |
| bloodio | Local-first web app that parses lab-report PDFs in the browser, validates and normalises values and charts trends; privacy by design, live demo |
| ai-radar | Weekly automated dashboard on open AI data: a Python pipeline on GitHub Actions collects arXiv, GitHub, job-ads and Eurostat indicators, a Svelte front end on GitHub Pages with a view on AI in Italy; every source licensed and cited (repository) |
| goal-line-calibration | Walk-forward evaluation of Elo, Dixon–Coles, logistic regression and XGBoost forecasts against de-margined bookmaker odds on 45,000 matches: frozen hyperparameters, paired bootstrap, calibration, published report |
Italian (native) · English (C2) · French (C1) · Spanish (C1)



