Geohash Converter — Encode and Decode Lat/Long

Real-world use cases

• Build a proximity search without a geo database

  You have a list of stores in plain SQL or a Redis sorted set and you want a cheap nearby query. Encode each location to a geohash, index it as a string, and a where-geohash-starts-with-u4pru filter returns everything in that cell. Use this tool to spot-check a few records by hand, confirm the prefix length matches the radius you care about, and grab the 8 neighbors so your border cases do not silently drop points.

• Debug coordinates a service returned to you

  An API logged a geohash but you need the actual point on a map. Paste the hash, read the decoded center and the bounding box, and you know both where it is and how exact it is. The plus-or-minus error tells you immediately whether the upstream system stored a city-level 5-char hash or a precise 9-char one, which is often the real bug.

• Pick the right precision for a feature

  Storing user check-ins? A 6-char geohash (about 1.2 km) protects rough privacy while still clustering a neighborhood; a 9-char hash pins a storefront. Slide the precision control and watch the cell size update so you choose a length on purpose instead of copying a magic number from a tutorial.

• Generate test fixtures for geo code

  Writing unit tests for a mapping feature? Encode a handful of known landmarks at a fixed precision, copy the strings into your fixtures, and you have stable, human-readable inputs. The shareable URL lets you drop a teammate straight onto the exact case you are reproducing.

Common pitfalls

• Swapping latitude and longitude. Latitude (-90 to 90) is the first field and longitude (-180 to 180) is the second. Feeding 116.4 as a latitude silently produces a wrong but valid-looking geohash, because the value just gets clamped to 90. Double-check which number is which before trusting the output.

• Relying on prefix matching alone for nearby queries. Two points a meter apart can land in different cells if they straddle a boundary, so a where-starts-with filter misses them. Always include the 8 neighbor cells and then filter by real distance, otherwise your radius search has blind seams along every grid line.

• Treating the decoded center as the original point. A geohash names a box, not a point, so decoding always returns the box center with a built-in error of half the cell. If you re-encode that center you get the same hash back, but it is not the exact coordinate you started from unless the precision is very high.

Privacy