Our data and analysis

We want to show you what data we collect, and how we analyze it.

Key numbers

Hours measured
—

Hours walked

  1. Manhattan
  2. Brooklyn
  3. Queens
  4. Bronx
  5. Staten Island

Predictors

We predict noise using simple and common-sense features

These are the predictors of the Rumble score, the 0 to 100 ranking the map draws: 90 means louder than 90% of modelled New York streets. See our methodology section for more context on the score.

Expected move in Rumble score, in points

The score is a percentile ranking. By default, the effects refer to a one-standard-deviation shift in a feature. For binary features (0/1), we show the effect of a change from 0 to 1.

Each effect is marked stat sig the 95% interval excludes zero not stat sig the 95% interval includes zero 95% confidence interval no effect

  1. More shops and offices nearby
  2. Closer to an arterial
  3. Closer to an expressway
  4. A bigger class of road
  5. A wider road
  6. Closer to a main road
  7. No homes on this block (0/1)
  8. A weekday, same hour (0/1)
  9. More 311 honking complaints
  10. More 311 noise complaints
  11. Closer to an elevated train
  12. Closer to a traffic light
  13. Closer to an intersection

Key Numbers

Some important numbers on things we collect during our walks

Rows come from different sensors, and different walks often measure different outcomes.

Counted across every walk

  1. Honks heard
  2. Minutes of polluted air · PM2.5 over 100 µg/m³
  3. Minutes over 70 dB(A): Like a vacuum cleaner or a subway train
  4. Minutes over 85 dB(A): Like a food blender or a jackhammer
  5. Minutes of heavy fumes:· VOCs over 1,000 ppb

Book a walk · $100 How the numbers are built Open the map