Shapefile vs GeoJSON: Which Format for Spatial Data?

What each format actually is

A shapefile is an old Esri format that, despite the singular name, is a set of files that must travel together. The geometry lives in a .shp file, the attribute table in a .dbf file (a dBASE table), the coordinate system in a .prj file, and an index in a .shx file. A .cpg file may record the text encoding. Because it is several files, a shapefile is almost always shared zipped.

GeoJSON is a single text file in JSON, defined by an open standard (RFC 7946). One file holds the geometry and the attributes together as human-readable text. By specification, GeoJSON coordinates are longitude and latitude on the WGS84 datum, which removes a whole class of coordinate-system confusion.

Side by side

• Files: shapefile is many files (.shp, .dbf, .prj, .shx, and more) zipped together; GeoJSON is one .geojson file.
• Readability: GeoJSON is plain text you can open and diff; a shapefile is binary.
• Coordinate system: a shapefile can be in any projection and carries it in the .prj; GeoJSON is longitude and latitude by the standard.
• Attribute names: the shapefile's .dbf caps field names at 10 characters and has old type limits; GeoJSON property names are unrestricted.
• Size: shapefiles are compact for large datasets; GeoJSON is larger as text but compresses well and streams easily on the web.
• Ecosystem: shapefile is the lingua franca of desktop GIS and government data; GeoJSON is native to web maps and modern tooling.

Coordinate systems, briefly

This is where the two formats differ most in practice. A shapefile might be in longitude and latitude, or in a projected system measured in metres such as UTM or a national grid. That matters because area and distance are only meaningful once you know the units. GeoJSON sidesteps the question by fixing longitude and latitude, but then any real area or distance has to be computed geodesically rather than from the raw coordinates.

A good tool reads the .prj to tell projected data from geographic, and computes real area and distance accordingly, in square kilometres and kilometres rather than raw coordinate units. If you are choosing a classification and reading areas off the result, make sure the numbers are real. See how to make a choropleth map for the full workflow.

Which should you use?

• Use GeoJSON for web maps, for sharing a single readable file, for modern data pipelines, and when longitude and latitude is what you have.
• Use shapefile when you exchange data with desktop GIS or government portals, when the dataset is large, or when a projected national coordinate system must be preserved exactly.
• In many projects you will convert between them. Keep the .prj when you export a shapefile so the coordinate system is not lost.

How to open either one

You do not need desktop software to look at either format. Groundtruth opens both in the browser: drag a zipped shapefile (with its .dbf, .prj, and .cpg) or a .geojson onto the map and it parses the file locally, reads the attributes, detects the coordinate system, and lets you map, measure, join a spreadsheet, and export. Nothing is uploaded, so sensitive boundaries stay private.