Methodology

Data sources

1. planning.data.gov.uk

Approximately 65 councils submit structured developer agreement data to the national planning data platform. This includes individual agreements, contribution amounts, transaction records, and statuses. We download the full CSV exports and aggregate transactions into annual summaries by council, financial year, and purpose category.

2. Exacom PFM

Some councils use Exacom's developer contributions management system, which includes a public-facing module. Where available, we extract structured data from these pages.

3. PDF extraction

For the remaining councils (~240), we download their Infrastructure Funding Statement PDFs, extract the text using pdfplumber, and use AI-assisted extraction to pull out structured financial data. We use Anthropic's Claude models for this, with a confidence scoring system: low-confidence extractions are re-processed with a more capable model and flagged for manual review.

Purpose taxonomy

Councils use hundreds of different labels for spending categories. We map these to six standardised categories: Affordable Housing, Education, Highways & Transport, Open Space & Recreation, Health, and Other. The mapping is documented and consistent across all data sources. See our purpose pages for data by category.

Quality assurance

We spot-check approximately 20 councils per import cycle by comparing our extracted data against the source PDFs. This catches systematic extraction errors. For planning.data.gov.uk data, we validate aggregated totals against the original transaction records.

If you spot an error, please report it. We take data accuracy seriously and will correct errors promptly.

Update frequency

IFS data is published annually (councils must publish by 31 December). We update our database after each annual reporting cycle. planning.data.gov.uk data is checked quarterly. Exacom data is checked monthly.

Limitations

Our data depends on what councils publish. Some councils report minimal detail in their IFS — headline totals only, with no breakdown by purpose. Some publish late or not at all. Where data is missing, we show this clearly rather than estimating.

Financial figures may not be directly comparable between councils. Accounting treatments vary: some councils report gross figures, others net. Some include in-kind contributions (like affordable housing delivered on-site), others report only cash. We present the data as reported by each council.