Understanding GeoHash: Spatial Indexing and Its Application in Proximity Queries

GeoHash is a spatial indexing technique that transforms two‑dimensional latitude‑longitude coordinates into a one‑dimensional string, enabling fast proximity queries by leveraging prefix matching of the generated codes.

By dividing the Earth’s latitude range (‑90 to 90) and longitude range (‑180 to 180) repeatedly in half, a binary representation is produced for each coordinate; the bits are then interleaved (odd bits for latitude, even bits for longitude) to form a combined binary string.

The combined binary string is encoded using a custom Base32 alphabet (0‑9, b‑z without a, i, l, o), yielding a GeoHash such as "wx4g" that represents a rectangular area whose size decreases with longer strings (e.g., 5‑character codes cover ~10 km², 6‑character codes ~0.34 km²).

Longer GeoHash strings provide higher precision (8 characters ≈ 19 m, 9 characters ≈ 2 m), so the appropriate length depends on data density and query requirements.

GeoHash uses a Peano (Z‑order) space‑filling curve to map 2‑D space to 1‑D; while simple, this curve can cause “boundary” anomalies where distant points share the same prefix, which can be mitigated by also checking the eight neighboring GeoHash cells.

For point data GeoHash works well, but for line or polygon data other spatial indexes like R‑trees are often more suitable.