Research · 4 min read
1% of Seattle’s Street Segments Hold 28% of Its Property Crime
Reviewing the busiest 1% of segments covers more than a quarter of recorded property crime, and the ranking barely moves year to year.
One percent of Seattle’s street segments carry more than a quarter of its property crime — and they have done so, largely unchanged, for eighteen straight years. That is not a curiosity. It is a screening rule: review the busiest blocks first, because the rest of the city is comparatively quiet.
Our analysis of 706,100 records; 16,579 matched segments, January 1, 2008–June 30, 2026.
Key takeaways
- The busiest 1% of Seattle’s street segments — 165 of them — carried 27.9% of 521,145 matched property-offense records from 2008 through mid-2026.
- Widening the review to the busiest 10% (1,657 segments) covers 64.3% of matched records; the busiest 20% (3,315) covers 77.1%.
- Between 64% and 69% of top-decile segments were top-decile again the next year, across all 18 adjacent-year comparisons.
- About 26% of records could not be matched to a segment, and the ranking is a screening map of where records recur — not a verdict about any address.
How concentrated is property crime at the street-segment level, and does the concentration persist?
Across 706,100 Seattle property-offense records from 2008 through mid-2026, 165 street segments — about 1% — carried 27.9% of records, the busiest 10% carried 64.3%, and the same segments stayed on top for 18 straight years. The concentration is sharp, durable, and gives a clear place to look first.
The numbers behind the answer
Selected measures only. Denominators and interpretation stay attached so the headline cannot stand alone.
27.9%
Records on the busiest 1% of segments
165 street segments carried 145,253 of 521,145 matched property-offense records.
64.3%
Busiest 10% of segments
1,657 segments carried 335,149 records; the busiest 20% (3,315) carry 77.1%.
0.64–0.69
Top-decile retention, year to year
Across 18 adjacent-year comparisons, roughly two thirds of top-decile segments were top-decile again the next year.
The one-in-a-hundred blocks carry a quarter of the crime
Pull 706,100 property-offense records spanning 2008 through mid-2026, assign each to a street segment using the block address, and rank. The busiest 1% — 165 segments — hold 27.9% of matched records. Widen to the busiest 5% (828 segments) and you cover 51.2%; the busiest 10% (1,657 segments) cover 64.3%; the busiest 20% (3,315) cover 77.1%. The median segment carries ten records across eighteen years; the busiest carry dozens.
That is sharper than the published block-level finding, because a single 100-block label can cover several very different segments. It is also more demanding: assignment relies on the block address rather than a precise point, and about 26% of records could not be matched to a segment. The ranking is a screening map of where records recur — not a verdict about any address.
Screening rule
Review the busiest 1% of segments first, then check what is actually happening at the property.
Evidence visual
Records covered by reviewing the busiest segments
Cumulative share of 521,145 matched records across 16,579 street segments, pooled over 2008–2026.
The same segments stay busy for eighteen straight years
Concentration would be far less useful if it reshuffled every year. It does not. Compare each year’s top decile with the next: between 64% and 69% of segments stayed in the top group across all 18 adjacent-year comparisons, and the rank correlation among active segments held between 0.57 and 0.65. The yearly top-decile share itself barely moves — roughly 47% to 56% across nearly two decades.
Whatever drives recurrence, it is not a one-year artifact. A plan built on this pattern should still be refreshed, but the baseline is unusually durable.
Evidence visual
Top-decile retention, selected year pairs
Share of each year’s top-decile segments still in the top decile the following year.
Review the busiest segments first, then look at the property
The practical rule is a coverage rule. If you are deciding where to spend limited attention, start with the busiest 165 segments — about one in a hundred — and you have touched more than a quarter of recorded property crime. Go to the busiest 1,657 and you have touched nearly two thirds.
A recurring segment is a reason to look more closely, not a conclusion about a property. Block labels are blurred, a long segment can span very different conditions, and nothing here normalizes for how many people, vehicles, or businesses pass through. Use the ranking to choose where to inspect, then verify against the property itself.
- Rank segments by pooled and recent record count.
- Start with the busiest 1% (about 165 segments).
- Check what offense mix makes up each segment’s count.
- Ask whether a venue, transit node, or mixed land use dominates the segment.
- Re-rank on a schedule and verify conditions at the property.
The concentration is sharper at segments than at blocks
This curve is the steepest in the library. Across 16,579 matched segments, the top 1% held 27.9% of 521,145 records, the top 5% held 51.2%, and the top 10% held 64.3% — against a median segment of 10 records and a 95th percentile of 114.1, for a Gini of 0.7347. The classic benchmark, the law of crime concentration, reports roughly half of city crime on about 5% of segments; this dataset runs just above it, and the original Minneapolis micro-place study is Sherman, Gartin, and Buerger.
The stability numbers carry the same message as the shape. Top-decile Jaccard similarity runs about 0.48 to 0.52 year over year, 12-month retention about 0.65 to 0.68, and rank correlation about 0.62 to 0.64 — the local echo of the 14-year Seattle trajectories, which found citywide declines driven by a small set of improving micro-places.
One number qualifies all of it. The address-match rate is 73.81%, so about a quarter of the 706,100 rows are placed by nearest-segment geometry rather than a parsed block, which can sharpen apparent segment concentration. That is a reason to read the curve as a strong description, which hot-spots policing evidence shows is actionable, rather than as a property verdict.
Context
The local concentration is a textbook result; the screening rule follows from it.
Questions property teams ask
What is a street segment here?
A unit in Seattle’s Street Network Database — generally the block face between two intersections. It is finer than the published 100-block label but still not an exact address.
Does a busy segment mean my property is at risk?
No. It identifies where published records recur. It is a reason to inspect current conditions, not a probability for a specific property.
How many segments would I have to review to cover most of the crime?
The busiest 1% (165 segments) cover 27.9% of matched records; the busiest 10% (1,657) cover 64.3%; the busiest 20% (3,315) cover 77.1%.
Does stability mean the same segments will stay busy?
No. Retention describes past top-decile membership. It is unusually durable baseline context, not a prediction.
Is street-segment concentration a well-established finding?
Yes. It is one of the most replicated results in criminology, from the 1989 Minneapolis micro-place study through multi-decade trajectory work in Seattle and elsewhere. This article measures the same pattern locally.
Our methods, limits, and sources
How we calculated this
This is original 911 Sentinel research — we gathered the records, ran every calculation below, and published the aggregate dataset.
We assigned accepted Seattle property-offense records (2008–mid-2026) to Street Network Database segments two ways — block-address parsing against SND ranges and geometric nearest-segment — and ranked segments by record count.
- We pulled 706,100 property-offense records and 24,344 analytical street segments.
- We matched records to segments by block address (73.8% matched) and checked against geometry.
- We counted records per segment and computed cumulative coverage, Gini, and Lorenz points.
- We compared each year’s top decile with the next year’s and computed active-subset rank correlations.
What this analysis cannot establish
- Records are published SPD offense records, not unique events, victims, or forecasts.
- Block addresses and coordinates are deliberately blurred; a segment can span mixed conditions.
- About a quarter of records could not be confidently matched to a segment.
- No exposure denominator normalizes for people, vehicles, businesses, or operating hours.
- Concentration is descriptive and does not explain why records recur.
Sources
The raw records come from the sources below; the study design, analysis, charts, and conclusions are our own.
- Seattle Police Department — SPD Crime Data: 2008-Present
- City of Seattle — Street Network Database (SND)
- Published aggregate result (JSON)
- Source manifest (JSON)
- David Weisburd (Criminology) — The Law of Crime Concentration and the Criminology of Place
- Lawrence W. Sherman, Patrick R. Gartin & Michael E. Buerger (Criminology) — Hot Spots of Predatory Crime: Routine Activities and the Criminology of Place
- David Weisburd, Shawn Bushway, Cynthia Lum & Sue-Ming Yang (Criminology) — Trajectories of Crime at Places: A Longitudinal Study of Street Segments in the City of Seattle
- Anthony A. Braga, Andrew V. Papachristos & David M. Hureau (Justice Quarterly) — The Effects of Hot Spots Policing on Crime: An Updated Systematic Review and Meta-Analysis
Related questions and practical guides
No obligation · free property walk
Check the busiest segments against the property
Use the coverage rule to prioritize a review, then compare the segment with access points, parking, lighting, camera views, and surrounding land use at the property.