Live ML app that predicts which GP practices will overspend on NHS prescribing. Upload a raw NHSBSA file, get at-risk practices ranked by predicted overspend. Random Forest (R²=0.95) built with scikit-learn and Streamlit.

End-to-end NHS prescribing data science project. Analyses £10.54B and 217M+ records to find where the NHS overspends and predict at-risk GP practices using Python, DuckDB and Random Forest. Includes a live dashboard and a live ML prediction app.

Live interactive dashboard exploring £10.54B of NHS primary care prescribing across 42 ICBs and 5,764 GP practices. Built with Streamlit, DuckDB and Plotly. Eight chapters covering cost, geography, overspending, prediction and seasonality.

SQL analytics project using NHS English Prescribing data, PostgreSQL, validation checks, reproducible outputs, and healthcare cost analysis.

PostgreSQL NHS prescribing cost intelligence warehouse with validated 54M-row fact table, reproducible SQL outputs and SVG visualisations