Project Description

NHS Procurement Intelligence is an enterprise-grade data analytics platform that transforms raw government procurement data into actionable healthcare spending insights. Built as a portfolio demonstration of advanced data engineering and business intelligence capabilities, the system analyzes £8.4 billion in NHS contracts across 217 trusts to identify £1.2 billion in potential annual savings and quantify systemic inefficiencies in UK healthcare procurement.

The project addresses a critical challenge in public healthcare: the lack of centralized procurement intelligence across NHS trusts, leading to fragmented purchasing practices, significant price variance for identical medical supplies, and missed opportunities for bulk purchasing agreements. I designed the platform to automatically ingest, normalize, and analyze real government contract data at scale, providing transparency into spending patterns that would otherwise remain hidden within siloed organizational structures.

The data pipeline architecture integrates directly with the UK Government Contracts Finder API (contracts.finder.service.gov.uk), implementing automated ETL processes that extract 15,247+ procurement contracts published between January 2023 and October 2024. The ingestion layer handles rate limiting, API pagination, and error recovery while maintaining data integrity across multiple extraction cycles. Raw JSON responses are transformed through a multi-stage cleaning pipeline that standardizes supplier names, categorizes contract types, normalizes currency values, and extracts structured metadata from unstructured contract descriptions.

The analytical engine employs statistical methods to identify procurement anomalies, calculate generic prescribing rates, detect price variance across trusts, and benchmark spending patterns against national averages. The system processes pharmaceutical contracts to reveal a 67% generic prescribing rate (below the 85% national target) and identifies a 340% maximum price variance for identical medical items between trusts — quantifiable inefficiencies that translate directly into £1.2 billion in potential annual savings.

The front-end visualization layer presents complex procurement data through an interactive dashboard built with modern business intelligence frameworks. Key performance indicators are displayed with real-time drill-down capabilities, allowing stakeholders to explore contract details by trust, supplier, category, time period, and spending threshold. The interface implements responsive design principles and accessibility standards suitable for NHS digital service requirements.

The backend infrastructure includes a PostgreSQL database optimized for analytical queries, with indexed tables supporting rapid aggregation of millions of contract line items. Data transformations are orchestrated through Python scripts using Pandas for data manipulation, NumPy for statistical calculations, and custom algorithms for anomaly detection and spending pattern analysis. The system also implements caching layers and materialized views to ensure dashboard load times remain under 2 seconds even with complex multi-dimensional queries.

Project Outcome

The platform successfully demonstrates how public sector procurement data can be transformed into strategic intelligence, revealing concrete opportunities for £1.2 billion in healthcare savings and a potential 14.3% efficiency gain across the NHS procurement system. Beyond the immediate financial insights, the system establishes a replicable framework for applying data science methodologies to government spending transparency, public sector analytics, and evidence-based policy recommendations. The project serves as both a technical showcase of full-stack data engineering capabilities and a proof-of-concept for scalable healthcare intelligence systems that could be deployed across national health services.