The safest place in England is Bury

We have spent the last few days building a map of recorded crime for every constituency in England and Wales. It is free, it is on the Crimes Weekly website, and you can look up your own seat in about four seconds.

The first version of it was a lie.

Not a small one, either. Sorted from lowest crime to highest, the table read: Bury. Rochdale. Oldham. Bolton. Stockport. If you had put that on screen — and we very nearly did — you would have told several million people in Greater Manchester that they live in the safest place in the country.

They do not. Greater Manchester Police did not publish the data.

The number that gives it away

Here is the bottom of the table, in recorded crimes per 1,000 residents per year:

Bolton South and Walkden      0.2
Bury South                    0.2
Oldham East and Saddleworth   0.2
Manchester Rusholme           0.2
Rochdale                      0.3
Manchester Central            0.6
                              ↓  nothing at all in this gap
North Cotswolds              27.0
Forest of Dean               29.2

Table: Crimes Weekly · GB-MP-03 · Source: police.uk street-level recorded crime, 12 months to July 2026 · Census 2021 population

The average for England and Wales is 99. A seat reporting 0.2 is not a quiet corner of the country. It is a force that filed nothing.

What makes it obvious is the gap. Twenty-seven constituencies sit at or below 0.6, then there is empty space all the way up to 27.0, and then the data resumes and behaves normally. Real variation does not leave a hole like that. Every one of those twenty-seven is Greater Manchester.

So they are greyed out on the map, and left out of the rankings, and the page says why. Grey means unknown. It does not mean safe.

The harder case, which is the one that should worry you

Gloucestershire is worse, because Gloucestershire is nearly believable.

It published some months and not others. Its seats come out at 27 to 36 per 1,000 — the bottom of the real range, sitting just below genuinely quiet rural seats in Dorset and Devon at 41 to 44. Low, but not absurd. Nothing in the numbers screams.

There is no clean line to draw there. Draw it too high and you delete real low-crime countryside; too low and you publish a county's partial returns as fact. So we have done neither: Gloucestershire's figures are shown, and the page tells you plainly that the county published intermittently and its seats should be read with caution.

We would rather show you an honest uncertainty than a confident number we do not believe.

That is the general problem with crime data, and it is not really a technical one. A gap in publication looks exactly like an absence of crime. One is a filing failure and the other is good news, and in a spreadsheet they are the same cell.

Scotland publishes nothing at all

Not "publishes less". Nothing.

England, Wales and Northern Ireland all file street-level crime to police.uk every month — the PSNI included. Police Scotland does not participate. There is no Scottish equivalent of the file this map is built from, so there is no Scottish part of the map, and there cannot be one.

Scotland gets periodic aggregate statistics instead. They will tell you a national total. They will never tell you what happened on your street, or let you compare your area with the one next door, or let anyone build the thing you are looking at.

Policing is devolved, so this is a decision for the Scottish Parliament and Police Scotland — no Westminster MP can grant it. If you live in Scotland and you would like what the rest of the UK already has, your MSPs are the people to ask. Ask for something specific: monthly street-level crime data, published openly, at the granularity England, Wales and Northern Ireland manage already.

Three more traps, quickly

The worst place in the country is an office. Cities of London and Westminster tops the table at 667 per 1,000 — nearly seven times the average. It is not a war zone. The rate divides offences by residents, and almost nobody lives there while a million people a day arrive to work, shop and drink. Its shoplifting rate alone is 77 per 1,000 residents. The same distortion applies, smaller, to every town-centre seat in the country.

One month tells you almost nothing. The first version of this map used a single month, and a single bad week in a small seat moved it a long way. It now uses twelve months. That one change did more for the map's honesty than anything else we did to it.

Recorded crime is not crime. It is what was reported and logged. A rise can mean more offences or better reporting, and the things people do not report — most fraud, most domestic abuse — barely appear at all.

Why we are telling you this rather than just publishing the map

Because the map is the least interesting part.

Anyone can colour in 575 constituencies. The judgement is in what you refuse to colour in: the twenty-seven seats we greyed out, the county we showed with a warning instead of hiding, and the nation we cannot show at all. Those decisions are the difference between a map that informs people and one that quietly misleads several million of them about where they live.

Crimes Weekly is a weekly programme about UK crime — appeals, wanted people, and cases that have been through the courts. These articles are for the part that will not fit in a thirty-minute video: where the numbers come from, what they leave out, and when they should not be trusted.

The map is at crimesweekly.com. Go and find your own seat. If it is grey, now you know why.

Crimes Weekly is an independent programme. It is not affiliated with, endorsed by or speaking for any police force. Crime figures are police.uk street-level data under the Open Government Licence; boundaries are ONS.