Recorded offences
38,013
Southwark, in numbers
Crime and safety
Southwark council area. Every figure carries its own date.
What the recorded figures show, and what they cannot tell us.
Recorded crime and population
A larger population can mean more recorded offences even when the number of offences per resident falls. Read the total and the rate together.
Southwark · 2025/26 · Territorial police-recorded offences excluding fraud; British Transport Police excluded by CSP source. Source classifications retained.
Offences per 1,000 residents
120.8
Population used for the rate
About 314,800
2024 mid-year population
The source rounds this population to the nearest 100. The rate shown is the official published rate.
Show the figures and workings
The rate is recorded offences divided by the population, multiplied by 1,000. Where the source rounds its population figure, we retain its published rate rather than recalculate it from that rounded figure. This counts offences, not the percentage of residents who were victims.
| Period | Offences | Population used | Per 1,000 residents |
|---|---|---|---|
| 2025/26 | 38,013 | About 314,800 (2024) | 120.8 |
These figures match the source area for this period. That does not establish a comparable history for today’s council boundaries.
Official source and method →Can we test the population theory over 30 years?
Only where the offence definitions, recording practices and area boundaries are comparable. Changes in reporting, police recording and the types of crime matter too. We do not join incompatible series or invent missing years.
Only the latest independently checked observation is displayed. A comparable long-run crime-and-population trend has not been established.
- 2002/03: Series starts after April 2002 recording-standard introduction; not comparable to pre-2002 records.
- 2013/14: Offence grouping revision; source back-series retained. Recording practice changes also affect comparisons.
- 2017/18: Burglary classification changed; broad total retained but detailed groups need review.
- 2023/24: Burglary subcategories and recording-rule changes require interpretation review.
A comparable 30-year crime-and-population trend has not been established for this area.
For example, if offences rose by 10% while population rose by 20%, offences per resident would fall by about 8%. This is an illustration, not a finding about this area.
People commuting or visiting can also affect offences recorded in a place. The resident population is a useful common denominator, but does not count everyone present.
What could help? Follow the types and locations of repeated problems, check residents’ experiences, and assess the response. Area population totals do not identify who commits offences.
Local issues and proposed responses →Police-recorded locations for Southwark council area. Locations are approximate; these are not all crimes or a measure of personal risk.
40,131 mapped crimes
September 2025 to August 2026. Compared with 40,244 in the previous 12 months. Anti-social behaviour is counted separately.
Mapped crime over time
September 2024 to August 2026
Southwark council area · mapped crimes
Reviewed figures: September 2024 to August 2026. This reviewed record covers fewer than 30 years, so the chart opens on all of it.
Shown on this chart: 3,471 mapped crimes in September 2024; 3,604 mapped crimes in August 2026. These are recorded counts; check the coverage notes before comparing periods.
The line joins consecutive published observations.
Hover or tap the chart for a value. Use the date slider with touch or arrow keys. Escape closes the chart label.
Monthly crime and anti-social behaviour
| Month | Mapped crime | Anti-social behaviour |
|---|---|---|
| September 2024 | 3,471 | 878 |
| October 2024 | 3,776 | 905 |
| November 2024 | 3,487 | 754 |
| December 2024 | 3,336 | 663 |
| January 2025 | 2,992 | 630 |
| February 2025 | 2,915 | 626 |
| March 2025 | 3,368 | 785 |
| April 2025 | 3,158 | 813 |
| May 2025 | 3,436 | 906 |
| June 2025 | 3,501 | 969 |
| July 2025 | 3,461 | 954 |
| August 2025 | 3,343 | 889 |
| September 2025 | 3,198 | 843 |
| October 2025 | 3,397 | 844 |
| November 2025 | 3,255 | 785 |
| December 2025 | 3,200 | 674 |
| January 2026 | 3,024 | 636 |
| February 2026 | 3,034 | 696 |
| March 2026 | 3,285 | 758 |
| April 2026 | 3,253 | 843 |
| May 2026 | 3,324 | 973 |
| June 2026 | 3,568 | 805 |
| July 2026 | 3,989 | 899 |
| August 2026 | 3,604 | 923 |
Crime categories in the latest complete year
| Category | Mapped crimes |
|---|---|
| violent crime | 10,369 |
| theft from the person | 7,218 |
| other theft | 5,175 |
| shoplifting | 3,329 |
| vehicle crime | 2,454 |
| burglary | 2,302 |
| public order | 2,143 |
| robbery | 2,038 |
| drugs | 1,887 |
| criminal damage arson | 1,315 |
| bicycle theft | 1,156 |
| other crime | 406 |
| possession of weapons | 339 |
What these numbers include
Approximate published locations, not every crime or personal risk.
Anti-social behaviour is separate. Transport police and unusable locations excluded.
An empty ward count means no mapped rows in the complete required force files; it does not prove that no crime occurred.
These are retrieved mapped records, not a guarantee of complete crime reporting. Approximate locations and unmapped incidents limit comparisons.
Missing ward-months invalidate totals and changes. Busy centres attract visitors, so residential population is not a measure of the number of people exposed.
Police data and method →