NYC Odor Complaint Map

Method

About the NYC Odor Complaint Map

The NYC Odor Complaint Map is a New York City demonstration in the Infrastructure Intelligence suite. It turns public NYC 311 odor complaints into a defensible, spatial hypothesis about where odor is actually being released.

Approach

Odor complaints record where a smell was noticed, not where it originated. Wind, temperature inversions and falling barometric pressure move and release sewer gas, so a complaint cluster is a downwind signature of an upstream source.

The platform joins each complaint to the weather conditions observed at that moment, projects an upwind search corridor, and accumulates corridors across many complaints. Areas where corridors overlap repeatedly, under differing conditions, become ranked candidate source zones for field verification.

Planned data sources

Not connected
  • NYC 311 sewer odor complaints

    NYC Open Data, geocoded points

  • Historical weather observations

    Open-Meteo hourly reanalysis

  • OpenStreetMap

    Standard basemap rendered at 50% opacity

  • Wastewater treatment facilities (14)

    NYC ArcGIS wastewater treatment plants layer — authoritative

  • Wastewater pump stations (40)

    Open Sewer Atlas NYC and public records — approximate demonstration data

Why pump stations matter

Engineering rationale

Treatment plants are the largest and best-documented wastewater odor sources, but they are not the only ones. Pump stations, force mains, wet wells, air-release structures and vents concentrate turbulence and headspace release along the collection system, often far from any plant. A screening model that only knows about plants will systematically miss those releases and will over-attribute complaints to the nearest plant.

Including pump stations widens the candidate set so that a complaint upwind of a wet well can be explained by that wet well instead. Both asset classes are screened with the same distance and wind-alignment methodology — asset type and location quality never adjust the score, so a pump station is never favoured or penalised for what it is.

Infrastructure data and provenance

Read before interpreting results

Potential odor sources are screened using complaint location, historical wind conditions, distance, and available wastewater infrastructure locations. Wastewater treatment facilities and pump stations can both be potential odor sources.

WRRF locations in this demonstration are sourced from NYC ArcGIS. An authoritative machine-readable NYC DEP pump-station GIS dataset has not been incorporated. Pump-station locations are therefore compiled on a best-effort basis from Open Sewer Atlas NYC and related public sewer-system information and may be incomplete or approximate.

For a utility deployment, authoritative utility GIS would replace these estimated pump-station locations and could also include force mains, wet wells, air-release structures, odor-control facilities, manholes, vents, and other relevant assets. More complete and accurate infrastructure locations would improve candidate-source screening.

Screening results identify infrastructure geographically and meteorologically consistent with observed complaint conditions. They do not establish the actual source of an odor.

Limitations

  • Complaint reporting is uneven and reflects population density and awareness.
  • Station weather is interpolated and may not match street-level conditions.
  • Results indicate candidate areas for inspection, not confirmed sources.
  • Pump-station locations are compiled from secondary public sources and are incomplete and approximate; they are included to demonstrate the method, not to audit any specific asset.
  • The asset inventory omits force mains, manholes, vents and odor-control structures that authoritative utility GIS would supply.

Learn more

Your Data Our Science
Your Data Our Science

This demonstration is built by Your Data Our Science. Reach out to discuss applying the method to your own collection system, complaint records or utility GIS.

contact@yourdataourscience.com

Stack

0.1.0
  • React
  • TypeScript
  • Tailwind CSS
  • Leaflet
  • OpenStreetMap
  • Supabase