How to Find the Projection of a GeoJSON File (When Coordinates Are Wrong)

You've loaded a GeoJSON file into your map, and instead of seeing your data in California, you're looking at a tiny speck near Africa—or nothing at all. Or maybe your polygons are microscopic dots near the equator.

Welcome to one of GIS's most frustrating quirks: GeoJSON files with the wrong projection.

This guide explains why this happens and how to fix it.

Quick Solution: Use Projection Finder

Projection Finder detects this problem automatically:

1. Upload your GeoJSON — Drag and drop the file
2. Detects coordinate mismatch — Identifies if coordinates look projected, not geographic
3. Suggests likely CRS — Based on coordinate ranges and patterns
4. Preview on map — Test different projections to find the right one
5. Export corrected GeoJSON — Reproject to proper WGS84

The GeoJSON Spec Problem

Here's the core issue:

There is no mechanism in GeoJSON to specify a different coordinate system. The spec assumes WGS84. Period.

But in practice, many tools export GeoJSON with projected coordinates:

• CAD exports that don't reproject
• GIS tools with the wrong settings
• Scripts that copy coordinates without transformation
• Legacy systems that predate the RFC 7946 spec

The result? Files that look like valid GeoJSON but have coordinates in meters (UTM, State Plane) instead of degrees (WGS84).

What This Looks Like

Correct GeoJSON (WGS84)

{
  "type": "Feature",
  "geometry": {
    "type": "Point",
    "coordinates": [-122.4194, 37.7749]
  }
}

Coordinates: [-122.4194, 37.7749] — San Francisco

Broken GeoJSON (Projected Coordinates)

{
  "type": "Feature",
  "geometry": {
    "type": "Point",
    "coordinates": [552474.23, 4182938.15]
  }
}

Coordinates: [552474.23, 4182938.15] — UTM meters, but interpreted as degrees

How to Identify the Problem

Symptom 1: Data Appears at 0°, 0°

Your features cluster near the Gulf of Guinea (where the equator meets the prime meridian). This happens when large projected numbers get truncated or wrapped.

Symptom 2: Features Are Microscopic

Your polygons appear as tiny dots. This happens when meter values (like 500,000) are interpreted as fractions of a degree.

Symptom 3: Nothing Shows at All

Some viewers reject coordinates outside valid lat/lon ranges (-180 to 180, -90 to 90). If your UTM easting is 552474, that's not a valid longitude, so the viewer shows nothing.

Quick Diagnostic: Look at the Coordinates

Open your GeoJSON in a text editor and find the coordinates arrays:

Coordinate Range	What It Means
[-180, 180] and [-90, 90]	Valid WGS84—this is correct
Large numbers (100,000+)	Projected coordinates—needs conversion
Numbers like [529090, 181440]	Likely British National Grid
Numbers like [552474, 4182938]	Likely UTM

Finding the Actual Projection

Once you've confirmed your GeoJSON has projected coordinates, you need to identify which projection.

Method 1: Use Projection Finder (Fastest)

1. Go to projectionfinder.com
2. Drop your GeoJSON file
3. Tool detects "coordinates don't look like WGS84"
4. Suggests likely projections based on coordinate patterns
5. Preview each option until the map shows the right location

Method 2: Context Clues

Ask yourself:

• Where is this data supposed to be? — If California, check UTM Zone 10N or California State Plane
• Where did the file come from? — CAD exports often use local coordinates; government data uses regional projections
• Is there accompanying documentation? — Metadata files or README might specify the CRS

Method 3: Coordinate Range Analysis

X Range	Y Range	Likely Projection
100K - 900K	0 - 10M	UTM (various zones)
0 - 700K	0 - 1.3M	British National Grid (EPSG:27700)
100K - 3M	0 - 500K	US State Plane (various zones)
-20M - 20M	-20M - 20M	Web Mercator (EPSG:3857)

How to Fix the GeoJSON

Once you know the source projection, you need to reproject the coordinates to WGS84.

Option 1: Use Projection Finder

Projection Finder can export your data as properly-projected WGS84 GeoJSON once you've identified the source CRS.

Option 2: QGIS

1. Open QGIS
2. Drag your GeoJSON file onto the map
3. When prompted for CRS, enter the source projection (e.g., EPSG:32610 for UTM Zone 10N)
4. Right-click the layer → Export → Save Features As
5. Set CRS to EPSG:4326 (WGS84)
6. Save as GeoJSON

Option 3: Python with GeoPandas

import geopandas as gpd

# Read GeoJSON with the ACTUAL source CRS
gdf = gpd.read_file('broken.geojson')
gdf = gdf.set_crs('EPSG:32610')  # Source CRS (e.g., UTM Zone 10N)

# Reproject to WGS84
gdf_wgs84 = gdf.to_crs('EPSG:4326')

# Save corrected GeoJSON
gdf_wgs84.to_file('fixed.geojson', driver='GeoJSON')

Option 4: ogr2ogr (Command Line)

# Reproject from UTM Zone 10N to WGS84
ogr2ogr -f GeoJSON \
  -s_srs EPSG:32610 \
  -t_srs EPSG:4326 \
  fixed.geojson \
  broken.geojson

What About the Old "crs" Property?

You might see GeoJSON files with a "crs" member:

{
  "type": "FeatureCollection",
  "crs": {
    "type": "name",
    "properties": {
      "name": "EPSG:32610"
    }
  },
  "features": [...]
}

If you have a GeoJSON with a crs property, don't trust that the coordinates actually match. Verify by previewing on a map.

Frequently Asked Questions

Summary

1. GeoJSON requires WGS84 — The spec mandates [lon, lat] in degrees
2. Many files violate this — Tools often export projected coordinates incorrectly
3. Identify the problem — Large numbers (100,000+) indicate projected coordinates
4. Find the source CRS — Use context clues or Projection Finder
5. Reproject to WGS84 — QGIS, Python, or ogr2ogr can transform coordinates

Related Guides

• How to Find the Projection of a Shapefile
• How to Find the Projection of a CSV File
• How to Find the Projection of an Excel File

Related Resources

• EPSG.io — Search and lookup EPSG codes
• spatialreference.org — CRS definitions in multiple formats
• I Hate Coordinate Systems — Troubleshooting guide for projection problems