Boroughly › Data & sources

Data & sources

What Boroughly's numbers actually are, where each one comes from, and which are still estimates. Updated whenever the data pipeline runs.

Read this before relying on a figure. Prices are median sale prices from HM Land Registry records — real transactions. Scores are still indicative composites. An area median is not a valuation of any particular home: two streets five minutes apart can differ by hundreds of thousands of pounds.

Status by dataset

City	Dataset	Status	What that means	Source
London	prices	Sourced	Median sale price per postcode district from HM Land Registry Price Paid Data, covering 278 of 278 areas. Areas with fewer than 30 recorded sales are left as estimates and flagged individually.	HM Land Registry Price Paid Data
London	crime	Estimated	Safety scores are composite judgements, not derived from police.uk recorded crime counts.	—
London	schools	Estimated	School scores are composite judgements, not derived from DfE performance tables.	—
London	transport	Sourced	London journey times are fetched live from the TfL Journey Planner API at request time.	Transport for London Unified API
Manchester	prices	No data	No data. Area names and coordinates only.	—
Manchester	crime	No data	No data.	—
Manchester	schools	No data	No data.	—
Manchester	transport	No data	No data.	—
Birmingham	prices	No data	No data. Area names and coordinates only.	—
Birmingham	crime	No data	No data.	—
Birmingham	schools	No data	No data.	—
Birmingham	transport	No data	No data.	—

What the statuses mean

Sourced

Derived directly from a named primary source by an automated pipeline, with the fetch date recorded. Reproducible — run the script and you get the same number.

Estimated

An indicative figure entered into the dataset rather than derived from records. Directionally useful for comparing areas, but not a measurement.

No data

The area exists in the system with a name and location only. Nothing is published for it.

Where the sourced data comes from

As each dataset moves from estimated to sourced, it is pulled from these, all of which are free and openly licensed:

HM Land Registry Price Paid Data — every residential sale in England and Wales. We take the median of the last 12 months per postcode district, never the mean, because a single high sale distorts an average badly. Open Government Licence v3.0.
police.uk Street-Level Crime API — recorded crime by location, converted to a rate per 1,000 residents so dense areas aren't unfairly penalised. Open Government Licence v3.0.
DfE school performance tables — the share of schools rated Good or Outstanding in each local authority.
ONS — population and boundary data used to turn raw counts into comparable rates.
Transport for London Unified API — live journey times, fetched when you use the map rather than stored.

How the scores are built

Where a score is sourced, it's a decile rank across every area we cover: a 10 for safety means the area sits in the lowest-crime tenth of our dataset, not that it is objectively safe. We publish the transformation rather than a judgement, so you can disagree with the weighting and still trust the underlying number.

What the data can't tell you

Averages hide enormous street-level variation — two roads five minutes apart can differ by hundreds of thousands of pounds. Crime statistics reflect what was reported, not what happened. School catchments move annually. Community sentiment scores are opinion, collected imperfectly. Every one of these is a reason to visit an area at different times of day before deciding, and none of them is a reason to skip that step.

Corrections

If a figure looks wrong, it may well be. Tell us and we'll check it against the source and publish a correction.