A geohash is a short string that encodes a latitude and longitude into a single base-32 token. Nearby places usually share a common prefix, which makes geohashes useful for spatial indexing and proximity queries.

How does precision relate to area?

Each extra character roughly cuts the cell size by a factor of eight, alternating between latitude and longitude refinement. A 6 character geohash covers about 1.2 by 0.6 km, while a 9 character one is a few metres across.

Why are some letters missing from a geohash?

Geohash uses a 32 symbol alphabet that omits a, i, l and o to avoid confusion with similar looking characters. The valid set is 0-9 plus b, c, d, e, f, g, h, j, k, m, n, p, q, r, s, t, u, v, w, x, y, z.

Do shared prefixes guarantee closeness?

A shared prefix means two points fall in the same cell, so they are close. The reverse is not always true: two nearby points can straddle a cell boundary and differ early in the string, the well-known edge case of geohashing.

Is the algorithm reversible?

Yes. A geohash decodes back to the bounding box and center of its cell. Use the companion Geohash to Lat/Lng tool to go the other direction.

Lat/Lng to Geohash Converter

Geohash encoding

A geohash turns a pair of coordinates into a compact, sortable string. It is widely used in databases and search engines because points that are near each other usually share a long common prefix, letting you find neighbours with a simple range scan.

How it works

The algorithm interleaves bits of longitude and latitude. Starting from the full ranges of -180 to 180 for longitude and -90 to 90 for latitude, it repeatedly halves a range and records a 1 if the coordinate falls in the upper half or a 0 otherwise. Bits alternate between longitude (first) and latitude. Every five bits become one base-32 character.

lat 57.64911, lng 10.40744  ->  u4pruydqqvj

The more characters you request, the more halving steps run, and the smaller the resulting cell.

Example and tips

Encoding a city-center coordinate at precision 6 gives a cell roughly a kilometre wide, which is handy for clustering. For pinpoint accuracy choose precision 9 or higher. Remember the boundary edge case: two physically close points may not share a prefix if they sit on opposite sides of a cell edge, so for robust neighbour search you should also check adjacent cells.