Make quick and dirty projections to use in your web maps instead of Web Mercator
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

Dirty Reprojectors

Quick and dirty re-projections to trick your web maps out of web mercator.


npm install -g dirty-reprojectors



cat input.geojson | dirty-reproject --forward PROJECTION [--reverse PROJECTION=mercator] > output.geojson

Example: to reproject some geojson so that web mapping libraries will render it looking like 'albersUsa':

cat input.geojson | dirty-reproject --forward albersUsa > output.geojson

For a list of supported projections, dirty-reproject --list



Reprojects the given geometry coordinate array in place, with unprojectable points or degenerate geometries removed. If both options.forward and options.reverse are supplied, then forward is performed first.


  • options Object
    • options.forward (Function | string)? The forward projection to use.
    • options.reverse (Function | string)? The reverse projection to use.
    • options.projections Object? A map of named projections to use. If provided, then string values of options.forward or options.reverse will be used as keys to look up the projection function in options.projections. For an extensive list provided by d3-geo-projection, use require('dirty-reprojectors/projections').
  • coordinates Array

How it works

Take, for example:

cat input.geojson | dirty-reproject --forward albersUsa > output.geojson

What this actually does is:

  1. Project input.geojson from WGS 84 (longitude/latitude) into albersUsa, with the target coordinates scaled to match the dimensions of Web Mercator.
  2. Reverse-project the result back to WGS84 as if it had been projected with Web Mercator. So now, when your favorite web mapping library tries to project it into mercator, the geometries end up looking like they were projected using Albers.

The main catch is that if you actually look at the longitude/latitude coordinates in output.geojson, they are totally wrong. (There are other, subtler catches, too, having to do with Web Mercator's limited latitude range, varying loss of precision, and probably many other nuances I am not aware of.)