What police.uk crime data actually means

Where the data comes from

police.uk is the official open-data service for recorded crime in England, Wales and Northern Ireland. Forces upload their data every month, and it’s published under the Open Government Licence for anyone to use. It’s the same feed sitting behind the crime layers on the big property portals; PostcodeLens uses it too, so the underlying numbers are the same — the difference is in how they’re framed.

What an “LSOA” is, in plain English

police.uk publishes street-level crime, but it groups the figures by a statistical building block called anLSOA — a Lower Layer Super Output Area. Ignore the mouthful: an LSOA is simply a small, official geographic area of roughly 1,500 residents (around 1,600 on average), designed by the Office for National Statistics to be a consistent unit for comparing places. There are about 35,000 of them across England and Wales.

One more wrinkle worth knowing: the map points police.uk shows are deliberately snapped to representative locations — the centre of a street, a public place — each covering at least eight addresses, so no incident is ever pinned to an individual home. The dot on the map is an anonymised approximation, not the actual address.

The 14 crime categories

police.uk sorts recorded crime into 14 categories, mapped from the Home Office’s more detailed offence codes. They are: anti-social behaviour, bicycle theft, burglary, criminal damage and arson, drugs, other theft, possession of weapons, public order, robbery, shoplifting, theft from the person, vehicle crime, violence and sexual offences, and a final “other crime” catch-all.

Two of these are worth reading with care. “Violence and sexual offences” is a single broad bucket that spans a very wide range of seriousness, so a high count there doesn’t by itself tell you what kind of incidents dominate. And “anti-social behaviour” is often the largest category in a town or city centre — it reflects nuisance and disorder, which is not the same as the acquisitive or violent crime most buyers are picturing.

Why small-area counts get suppressed — and why that isn’t zero

When you narrow crime down to a very small area, the counts get small too — and small counts create a privacy problem, because they can edge towards identifying the people involved. For that reason, PostcodeLens suppresses a crime figure when there are too few recorded incidents to publish safely, rather than displaying a precise but privacy-eroding number.

The crucial point — and one PostcodeLens states outright on the page — is that suppression is not the same as zero crime. When we widen the area as far as the local district and there are still too few incidents to disclose, the page says so plainly:“This does not mean crime is low here; it means the count is too small to disclose safely.” A blank where a number would go is a statement about privacy thresholds, not about safety.

When a postcode does have enough incidents to report honestly, PostcodeLens starts at the immediate neighbourhood (the LSOA) and only widens — to the local ward, then the district — if the finer area is too sparse, always telling you which area the figure covers. A worked example: compare a dense city-centre postcode likeM1 2HF in Manchester withBS1 5TR in Bristol.

Why a raw count on its own can mislead

A single number — “480 crimes in the last 12 months” — is almost meaningless without something to compare it to. A busy city-centre postcode packed with shops, bars and thousands of daytime visitors will record far more crime than a quiet rural one, simply because there are vastly more people and premises for crime to happen to. That doesn’t automatically make the city centre “more dangerous” for a resident; it makes it busier.

The honest way to read a count is like for like: compare similar places (city centre to city centre, suburb to suburb), look at the mix of categories rather than the headline total, and be wary of any tool that ranks areas on raw counts alone without accounting for how many people and businesses are actually there.

The lag, and “recorded” crime

Two final caveats that apply to all police.uk data, wherever you read it:

• There’s a publication lag. Forces upload data after each month closes and it goes through quality assurance, so the most recent couple of months are typically not yet available. Crime data always describes the recent past, never today.
• It’s recorded crime, not all crime. The figures count offences that were reported to and logged by the police. Crime that is never reported — and reporting rates differ a lot between offence types — simply doesn’t appear. The data is a measure of what police recorded, which is related to, but not the same as, everything that actually happened.

In short

• All the numbers come from police.uk, grouped by small ~1,500-resident LSOAs.
• There are 14 categories — read the broad ones (violence & sexual offences, anti-social behaviour) with care.
• A suppressed figure means “too few to publish safely”, not “no crime”.
• Raw counts need a like-for-like comparison; dense places record more.
• Expect a couple of months’ lag, and remember it’s recorded crime, not all crime.

Next: how sold house-price data really works, or how to check flood risk before buying. For how we aggregate and attribute figures, see themethodology page.