H3 res-9 hexagon surfaces from Rentometer rents or your own sales comps. Runs entirely in your browser; your API key and data never touch a server of ours.
Click the map to place vertices. Double-click (or press Enter) to finish; Esc cancels.
Polygon/MultiPolygon layers only. Multiple features can be dissolved into one study area. Reprojection to WGS84 is automatic.
Use these to drop lot sales, mobile homes, and data errors before they smear into the surface — check the min/median/max readout below.
Each res-9 hex gets an inverse-distance-weighted (power 2) average of its k nearest comps within the search radius; hexes with no comps in range stay null. low_confidence = fewer than 3 comps used or nearest comp > 0.5 mi. No API credits involved.
"Full" guarantees no gaps along the study-area edge but extends one hex past the boundary in places (slightly more API calls). "Centers only" keeps every hex center inside the boundary.
When sampling is coarser than res 9, downfilled hexes flat-copy their parent, which reads as a patchwork quilt. Smoothing interpolates each res-9 hex from nearby sample centers instead (inverse-distance-weighted, power 2). Raw values are kept; smoothed values go in a separate percentile_75_smoothed column and drive the map display. No effect when sampling at res 9.
If set, the output adds percentile_75_per_sqft and median_per_sqft = rent ÷ this value. It's a uniform divisor — handy for $/sqft units, but it does not capture how unit sizes vary across the study area (Rentometer's QuickView endpoint returns no sqft data).
The metric of record is always percentile_75; all other summary stats are captured alongside it.
Key is kept in this browser tab only (sessionStorage) and sent directly to rentometer.com. The check itself is free (/auth, 0 credits).
The combined file codes both metrics onto one res-9 layer (joined on hex id) and adds gross_yield_proxy = 12 × rent $/sqft ÷ sales $/sqft where both exist (needs an assumed unit size on the rent run). Switching modes keeps each run; changing the study area clears both.
Every /summary call costs 1 QuickView credit, including repeats — there is no cache. The credit gate above is the only guard rail. Tip: a previously downloaded GeoPackage from this tool documents exactly what was measured (see sample_points + run_metadata).
