Recorded crime in BA20 is 2.0 times the Somerset average and 1.6 times the England and Wales average. Over 5 years it has risen 11%. Most of the harm comes from violence and sexual offences (assaults, harassment, sexual offences). Police data to August 2026.
Yeovil, Somerset. 12,852 residents in 9 neighbourhoods.
What it measures. How serious the crime recorded here is, not how much of it there is. Each recorded crime is weighted by the typical sentence for that kind of offence (the Office for National Statistics Crime Severity Score), so a robbery counts about 76 times a shoplifting offence. The total per resident is ranked against every neighbourhood. 10 means among the safest tenth.
Data. Street-level crime from data.police.uk, September 2016 to August 2026. Weights: ONS Crime Severity Score, year ending March 2025.
Crime severity: 2.0× the Somerset average (1.6× the England and Wales average)
Crime only, by how serious it is. Compared with every postcode district. Higher is safer.
What it measures. One score from 1 to 10 that blends schools (33%), transport (25%), health (25%) and supermarkets (17%). Crime is not in it: it is the separate Safety score. 10 is best.
Data. Built from the four scores below, so it uses their data. Figures last rebuilt on 10 October 2026.
Based on schools, transport, health, supermarkets, weighted by importance. Crime is not part of it. Higher is better.
1 neighbourhood has no population figure, so it is not scored.
Up 11% in 5 years Recorded crimes in BA20, anti-social behaviour left out.
Crime severity: 2.0× the Somerset average (1.6× the England and Wales average). 145 recorded crimes per 1,000 residents a year.
"Violence and sexual offences" is 66% of the harm here. It is a broad police category: most of these crimes are assaults without injury, harassment and threats, often indoors between people who know each other, and only a small part are sexual or serious violent offences. Police data does not say which, so each one is weighted at the England and Wales average for the category, which can overstate a quiet residential area.
Anti-social behaviour: 39 incidents per 1,000 residents a year, not scored. Not a crime, so it is not part of the Safety score.
Distances are straight-line, not walking routes. Scores use fixed distance bands, so they do not depend on other districts. Higher is better.
What it measures. Recorded crimes per 1,000 residents a year, with burglary and robbery counting more than shoplifting, ranked against every neighbourhood. 10 means among the safest tenth.
Data. Street-level crime from data.police.uk, September 2016 to August 2026.
Recorded crimes per 1,000 residents per year, most serious first, not counting anti-social behaviour.
| Type | Per 1,000 a year |
|---|---|
| Robbery | 1.9 |
| Burglary | 5.4 |
| Violence and sexual offences (assaults, harassment, sexual offences) | 52.8 |
| Drugs | 2.6 |
| Possession of weapons (knives and other weapons) | 1.2 |
| Other crime | 2.0 |
| Other theft | 12.1 |
| Theft from the person (pickpocketing, bag snatching) | 0.8 |
| Vehicle crime (cars stolen and broken into) | 3.9 |
| Criminal damage and arson (vandalism) | 14.3 |
| Public order | 23.7 |
| Bicycle theft | 1.1 |
| Shoplifting | 23.4 |
Crime severity per resident (sentence-days of the crimes recorded each year). Hatched bars are years with fewer than twelve months of data for most residents.
Last 12 months: 29.4 against a 10-year average of 23.7, 24% higher.
Council: Somerset
Few people live in these areas but many shop, work and go out there, so crime per resident looks far higher than the risk to someone living there. They have no Safety score, because compared with residential areas the number would mislead; their crime figures are still shown.