We would rather publish fewer figures we can defend than more we cannot. This is exactly how the data gets sourced, aggregated, and reviewed before it reaches a page.
Active and recently-sold listings are sourced from Cotality, the authorized distributor of REBNY's Residential Listing Service (RLS). Every listing carries its originating brokerage attribution.
Demographic and housing-stock layers are joined from the US Census American Community Survey (ACS) 5-year estimates on standardized NTA GEOIDs.
Mortgage rates are pulled from the Freddie Mac PMMS weekly survey — never a vendor-blended composite.
Every headline figure is a median-of-medians by Neighborhood Tabulation Area (NTA), then weighted by reporting inventory at the borough level. A single super-tower cannot skew a neighborhood line.
Scanner rankings (newest, most-expensive, biggest cuts, longest on market) are deterministic SQL over the live feed — no editorial thumb on the scale, and the ordering rule is published on each scanner page.
Listing surfaces refresh on a 6-hour cadence from the live feed. The quarterly Denza Index and the Open Datasets are versioned snapshots published with a fixed methodology tag so historical figures never silently change.
Each scanner page discloses its own refresh cadence so a reader always knows how fresh the underlying data is.
All published datasets are aggregate. No individual listing keys, no agent PII, no addresses below the building level. Every release passes a REBNY Data License Agreement (DLA) review before it ships.
We publish no discriminatory filters of any kind — see our Fair Housing commitment for the full statement.
Rentals are excluded from sale-price aggregates. Off-market and pocket inventory are not represented in public figures. Co-op buy-side comps reflect the last recorded sale, not the board-approved transfer price.
Where a figure is modeled rather than observed (e.g. affordability ratios), the model inputs are listed in the dataset schema.