A transparent methodology for public-sector project intelligence

Public-only source discovery

Sources are discovered through official websites, sitemaps, public document libraries and verified public-sector domains. Crawling respects access controls, robots rules, response limits and configured collection boundaries.

Extraction and classification

Documents are parsed before relevant fields are extracted. A useful record needs an identifiable organisation, credible source and evidence connected to a potential built-environment requirement. Vague aspiration alone is not enough.

Evidence, inference and confidence

Evidence is what the source states. Inference is the commercial interpretation of what that evidence may mean. Confidence can increase when dates, named sites, approvals and repeated independent references align, but automated analysis can still be wrong or superseded.

• Evidence remains linked to its original source.
• Inferences are presented as judgement, not fact.
• Material opportunities should be checked against later documents and official notices.

Corrections and limitations

Projects can be delayed, cancelled, renamed or absorbed into wider programmes. Teams should report material corrections and avoid relying on a single stale document. Project Pulse does not claim to identify every project or guarantee a procurement outcome.