DROP · VANDALISMSEPTEMBER 2026 BRIEFINGSAN FRANCISCO · 25.8K residents

South of Market Crime Rate Trends — San Francisco

South of Market — universally called SoMa — is the formerly industrial district south of Market Street, transformed since the 1990s into a dense mix of tech offices, condos, museums (including SFMOMA), and the convention space at Moscone Center. It carries a 24-hour rhythm distinct from the rest of the city, with nightlife clusters, the Caltrain terminus, and an uneven mix of long-time residents, single-room-occupancy hotels, and new-build towers.

VANDALISM · 24-MO COUNT09 2026 · 38
0336612-mo avg: 39.0
SOUTH OF MARKETCITYWIDE TREND (RESCALED)-20% 12MO YOY
+15%MoM
-20%12mo YoY
468last 12mo
38this month
01 · TL;DR

Four signals surfaced in South of Market this September, split evenly between one-month below-trend moves and structural sustained shifts. Two categories, vandalism and burglary, ran below their recent levels in a single-month drop. Burglary also registered as a sustained shift, meaning the decline is multi-month, not just a quiet September. The overall shape is a narrowing property crime footprint, with the longer-term compression in burglary the clearest structural story.

Burglary is down 33.2% over the trailing 12 months, 245 incidents against 367 in the prior year, and the sustained-shift signal reinforces that this is not noise. Vandalism follows a similar arc, off 19.7% year-over-year at 468 incidents vs. 583. Outside property crime, the picture is mixed: theft from vehicle is down 38.5% on a 12-month basis, while other larceny is up 9.6% and sexual assault has risen from 10 to 15 incidents year-over-year. Everything else remained within expected range.

2 drops2 sustained shifts
02 · Notable signals

Notable signals 2

DROP · VANDALISM

Vandalism

The past 12 months saw 468 incidents — about 34% below the 711 average from prior years.

DROP · BURGLARY

Burglary

The past 12 months saw 245 incidents — about 52% below the 515 average from prior years.

03 · By category

All categories, last 24 months

Each panel: recent monthly count vs. trailing 12-month context. MoM is the most recent month vs. the one before; 12mo YoY compares the trailing year to the year before that.

Homicidebelow threshold
2024-102026-09
Robbery-5%
2024-102026-09
Aggravated Assault-6%
2024-102026-09
Sexual Assault+50%
2024-102026-09
Burglary-33%
2024-102026-09
Theft from Vehicle-39%
2024-102026-09
Other Larceny+10%
2024-102026-09
Motor Vehicle Theft-8%
2024-102026-09
Vandalism-20%
2024-102026-09
Arson+22%
2024-102026-09
05 · Forecast

What next month likely looks like

Forecasts trained through September 2026, with a likely range we're 95% confident the actual count will fall inside. Categories with too little recent volume — or violent categories at the neighborhood level — show no forecast and are surfaced through signals above instead. See the methodology page for the gating rules.

Aggravated Assault

NO FORECAST

Too low-volume per neighborhood for a reliable point forecast — see the rare-event and streak-break signals above instead.

Arson

OCTOBER 2026
Most likely 2 next month — likely between 0 and 5.
−24% vs 12-month average (≈2.3)

Burglary

OCTOBER 2026
Most likely 21 next month — likely between 0 and 41.
+4% vs 12-month average (≈20.4)

Homicide

NO FORECAST

Too low-volume per neighborhood for a reliable point forecast — see the rare-event and streak-break signals above instead.

Motor Vehicle Theft

OCTOBER 2026
Most likely 31 next month — likely between 14 and 47.
+22% vs 12-month average (≈25.2)

Other Larceny

OCTOBER 2026
Most likely 135 next month — likely between 41 and 235.
−4% vs 12-month average (≈141.3)

Robbery

NO FORECAST

Too low-volume per neighborhood for a reliable point forecast — see the rare-event and streak-break signals above instead.

Sexual Assault

NO FORECAST

Too low-volume per neighborhood for a reliable point forecast — see the rare-event and streak-break signals above instead.

Theft from Vehicle

OCTOBER 2026
Most likely 25 next month — likely between 0 and 69.
−7% vs 12-month average (≈26.9)

Vandalism

OCTOBER 2026
Most likely 41 next month — likely between 26 and 57.
+5% vs 12-month average (≈39.0)
06 · Context & comps

How South of Market compares

Peer neighborhoods picked by closest 12-month vandalism volume — a pragmatic v1 of peer matching. Demographic / housing-stock peer matching isn't built yet (we deliberately don't ingest income or race data alongside crime). Volume similarity has the right intuition: “neighborhoods experiencing comparable vandalism levels.”

07 · Patterns

Recurring local terms (last 12 months)

Top terms in incident descriptions for South of Market, excluding generic crime taxonomy. Useful as texture — what kinds of specifics show up here that don't show up elsewhere.

warrantpossessionnarcoticsparaphernaliashopliftingsaleofficeraggravatedtrafficmethamphetamineforceweaponpeacerecoveredinfluenceinvestigationcaselostresistingobstructingdelayingdutiessubstancecontrolledfound
When does it happen?

Hour-of-day, day-of-week, and seasonality

Distribution of bucketed incidents in this neighborhood across the full analysis window. Useful for routine context — shopping-strip thefts vs. late-night assaults read very differently when you can see when each typically happens.

HOUR OF DAY · ALL CATEGORIES
01,4202,83912am6am12pm6pm11pm

Hour 0 is mildly inflated by reports without a known time defaulting to midnight — see methodology.

DAY OF WEEK · ALL CATEGORIES
03,5487,095MonTueWedThuFriSatSun
MONTH OF YEAR · ALL CATEGORIES
02,0584,115JanFebMarAprMayJunJulAugSepOctNovDec
08 · Methodology

How we built this page

Data → Anomalies → Forecast → Page

Incident data is pulled from SFPD's open dataset on DataSF, mapped to 10 NIBRS-aligned categories, and aggregated to neighborhood × category × month.Anomalies are surfaced using strict thresholds (~p < 0.01). Forecasts are Prophet with low-count gating; violent categories at the neighborhood level skip the forecast and show rare-event / streak signals instead.

Spike rule: 12-mo total > baseline mean + 2.5σ AND ≥ 20 incidents AND 6-mo confirms. Drop rule: 12-mo total < baseline mean − 2.5σ AND baseline mean ≥ 20. Rare event: any incident in the last 90 days, no prior comparable in ≥ 5 years.