How a street gets its grade
Two numbers, each a rank. One for what happens on the block, one for what sits around it. Blended, cut into letters, placed among every graded street in the city. Here is Bedford Avenue in Brooklyn going through it.
Bedford Avenue, Brooklyn
Calm 65% of the grade
Each is complaints or crashes per home, ranked in the band
33Access 35% of the grade
Within a 400 m walk, one ladder for the whole city
79A 90 for quietness means quieter than 90% of the streets it can fairly be compared with. A 72 for transit means the same thing on every street in the city.
Result
0.65 × 33 + 0.35 × 79 = 49.1
2,848 of 5,856 graded streets score higher. Bedford Avenue sits at the 45th percentile.
Bedford Avenue collects sanitation complaints and housing violations at nearly the worst rate of any major corridor. Its best access measure is tree cover, at 100 out of 100 against every graded street in the city. The two halves pull apart, and the grade says so.
What each measure weighs
- Traffic safety 15.5%
- Street crime 9.8%
- Quiet 11.4%
- Clean 8.5%
- Housing 8.5%
- Infra 5.7%
- Parking 5.7%
- Transit 17.5%
- Food 10.5%
- Trees 7.0%
Traffic safety is the largest single weight because it is the one measure built from harm that already happened rather than from a report someone filed. It counts people hurt in NYPD collision records from 2021 on, a death at five times an injury. Street crime is the second such measure: felony assaults, robberies, murders and home break-ins the NYPD placed on the street, and it is the only crime number we publish.
Ranked against its own kind
A cul-de-sac and a six-lane avenue both collect noise complaints, and comparing their raw rates only tells you which one is the avenue. Each street is sorted by its type and how much is built on it, never by neighborhood, and every Calm measure is a rank inside that band. Access is one ladder for the whole city instead, because a train is a train everywhere. That makes Access a dial for how much the score is a map of Manhattan: 10% of graded streets, 250 of the top 500 at the current 35%. At 50% that share passes 59%. It stays at 35.
What the letter means
The score is built from two ranks, so it centers near 50 and always will. The letters are positions, not marks out of a hundred: a B at 60 is a street in the best quarter of everything we grade. The share under each letter is live. The cream tick is Bedford Avenue, the thin line is the median street at 49. The percentile a street page prints is that street's place on this same ruler, computed from the banded score, so a corridor's percentile is still earned against corridors.
Where it can mislead you
Traffic safety is crashes. Street crime is its own measure.
The two are built from two different NYPD files and never mixed. A high traffic-safety score means fewer people were hurt in crashes here than on streets of its kind. A high Street crime score means fewer felonies were placed here, and only the ones the NYPD could put midblock.
311 measures who calls.
Dividing by homes fixes size, not temperament. A block of non-callers can look calmer than it is. Ranking inside a band of similar streets softens this and does not remove it.
Two inputs are old.
The tree census is from 2015 and the city publishes crashes about ten weeks late. Where the numbers come from shows every input's date.
Streets move a little every week.
Three runs are on file. Trend arrows appear at four, and until then no page claims a direction.
Streets we do not grade
A street needs 20 homes before anything is divided by its residents. The floor was measured, not chosen: complaints per home, straight off the raw tables.
| Homes on the street | Streets | Complaints per home |
|---|---|---|
| 1 to 4 below the floor | 93 | 15.9 |
| 5 to 9 below the floor | 171 | 3.2 |
| 10 to 19 below the floor | 406 | 1.7 |
| 20 to 49 | 1,100 | 0.8 |
| 50 to 199 | 1,913 | 0.7 |
| 200 or more | 2,844 | 0.8 |
Measured 21 August 2026 against production. A design measurement, not a live count.
Above 20 the rate is flat. Below it the rate climbs twentyfold, because complaints about a three-flat street are filed by people who do not live on it. 1,074 streets fall below the floor (671 with one to 19 homes, 171 with none in the tax lot file, 232 with no housing record we could match), keep a page and their Access score, and carry no letter.
What changed and when
| Version | Dates | What it counted | What was wrong, measured |
|---|---|---|---|
| v1 | to Jun 2026 | Complaints, raw | 18% of scored streets sat on exactly 84 and another 12% on 77. |
| v2 | Jun to Aug 2026 | Complaints per kilometer, ranked in the borough | Length is not exposure. Streets with almost nobody on them rose to the top. |
| v3 live | Aug 2026 on | Per home, ranked in the band, blended with citywide Access | Safety reads 299.8K crashes instead of 56K nuisance calls. The top of the ranking is places people live. |
v1 and v2 figures measured 21 August 2026. v3 figures are live.