I build data pipelines end to end — from raw public-sector and regulatory sources to dimensional models and Power BI reports, across local, lakehouse and cloud warehouse environments.
| Project | Tools | Description |
|---|---|---|
| 📊 SEC EDGAR Financial Statements Pipeline | Python, DuckDB, dbt, SQL, Power BI | End-to-end pipeline that turns 18 million rows of raw SEC XBRL filings into a tested star schema for financial analysis. Every double-counting hazard in the source is an explicit, tested, documented flag, and every figure on the dashboard is checked against a written expectation before it ships. Runs on a laptop: no cloud account, no credentials. |
| 🚢 EU Trade Lakehouse | Python, PySpark, Delta Lake, Databricks, SQL | Medallion lakehouse over Eurostat's detailed EU trade data: seven months of 2026, 35 million rows, through bronze, silver and gold written by hand in PySpark on Databricks serverless. One Job lands each month from Eurostat and notices when a month is republished under the same file name; country and product codes keep their history as SCD type 2 dimensions, loaded by a hand-written MERGE; and the run stops on any of its 20 quality checks that finds other than the failures it declares. |
| Project | Tools | Description |
|---|---|---|
| 📈 US Restaurant Sector — Financial Analysis | SQL, DuckDB, Python, Power BI | Takes an evidence-based position on the relative financial health of ten listed restaurant companies: Yum China and Chipotle lead, and Bloomin' and Papa John's trail, whichever ranking pillar is left out. It is built on the warehouse above and reproduces that project's published figures before trusting a ratio of its own. Every figure names the query or filing behind it. |
| Project | Tools | Description |
|---|---|---|
| 🎨 Power BI Theme Lab | Node.js, Power BI, WCAG | A complete Power BI theme compiled from a short list of design tokens, with every colour pair it produces audited against the WCAG contrast formula. The live gallery shows the same report in six themes, where every figure is identical and only the palette moves. |
Described before they are built, because writing down what a project has to demonstrate is what keeps its scope honest. Listed without links until the repository exists.
| Project | Tools | Description |
|---|---|---|
| 🏛 EU Public Procurement Pipeline | Airflow, GCS, BigQuery, dbt, Python, XML, Power BI | Orchestrated cloud pipeline over EU tender notices for public spend analytics. |
Only what is actually demonstrated in a repository above. This section grows as the projects do.
| Category | Tools |
|---|---|
| Languages | Python, SQL, Node.js |
| Transformation | dbt, PySpark |
| Storage and formats | DuckDB, Parquet, Delta Lake |
| Platforms | Databricks |
| BI and reporting | Power BI, DAX, Power Query |
| Practices | Dimensional Modeling, Medallion Architecture, Data Quality, Incremental Ingestion, Financial Statement Analysis, Design Systems, Accessibility (WCAG) |
| Tooling | Git, GitHub Actions |