Uber H3 Hexagonal Index (beta)
Open BetabetaH3 is Uber's hierarchical hexagonal spatial index. Each map feature is aggregated into an H3 cell at the requested resolution; the export returns JSON with real H3 indices once the worker image is rebuilt with h3>=4.0. Until then the exporter fails honestly — we no longer emit fallback_7_7690_5336 synthetic strings.
Key Features
About H3 Format
H3 is a hexagonal indexing system for the Earth invented at Uber: the planet is split into hexagonal cells across fifteen levels of nesting, and every cell has its own short identifier. Hexagons are more convenient than squares in analytics: a cell has six neighbors at the same distance, so density is not distorted along the diagonals.
The export is what you need when the features themselves matter less than their distribution: how many cafes fall into a cell, where development thickens, how to compare districts with each other. The H3 index is the common denominator to which any map layer and your own data are reduced alike.
The output is a single .json with the H3IndexCollection structure: the resolution, the extent of the area and a list of cells. Every cell carries a real H3 index, the total number of features that fell into it and a breakdown by layer; on request the cell's boundary is added as a GeoJSON polygon. The resolution is set from 0 to 15, with seven by default — a cell with an edge of about 1.2 km, convenient at city scale.
The export is built honestly: indexes are computed by the h3 library version four and newer. If it is missing from the environment, the export returns an error rather than made-up placeholder strings — that mode existed before and was misleading. Features are assigned to cells by their centroids, and the limit is 500,000 features per export.
What Program Opens H3 Files
The following programs and applications can open, edit, and work with Uber H3 Hexagonal Index (beta) (.json) files exported from OSM2CDR:
Format specifications
- Open standard
- 2D only
- Lossy compression
- Stores feature attributes
- Output: 1 file
- Uber H3 hexagonal index aggregation as a single JSON file. Each record is an H3 cell ID + feature count (optional GeoJSON hex boundary). Resolutions 0-15, default 7.
Who Uses H3 Maps
How to Export H3
How to Open After Export
- Python — read the JSON, then use the h3-py library: h3.cell_to_boundary gives the cell outline, h3.grid_disk gives the neighbors, and pandas assembles the summary tables.
- deck.gl — the H3HexagonLayer takes indexes directly, so a density map is built without creating polygons first.
- Kepler.gl — load a table with a column of indexes and pick the H3 layer type: an interactive heat map comes together in a couple of clicks.
- QGIS (free) — if the export was made with cell boundaries, the polygons from the file attach as a layer and are colored by feature count.
- Databases — the H3 index makes a convenient join key against your own tables: ClickHouse has H3 functions built in, and PostgreSQL has the h3-pg extension.
Popular maps in H3
Get Uber H3 Hexagonal Index (beta) map exports
Download professional H3 maps from OpenStreetMap. Instant generation, any area.