django-geojson is a collection of helpers to (de)serialize (Geo)Django objects into GeoJSON.


django-geojson is a set of tools to manipulate GeoJSON with Django :

  • (De)Serializer for (Geo)Django objects, querysets and lists
  • Base views to serve GeoJSON map layers from models
  • GeoJSON model and form fields to avoid spatial database backends (compatible with django-leaflet for map widgets)


pip install django-geojson

This package has optional extra dependencies.

If you need GeoJSON fields with map widgets :

pip install "django-geojson [field]"


Add djgeojson to your applications :



(not required for views)

GeoJSON layer view

Very useful for web mapping :

from djgeojson.views import GeoJSONLayerView
url(r'^data.geojson$', GeoJSONLayerView.as_view(model=MushroomSpot), name='data'),

Consume the vector layer as usual, for example, with Leaflet loaded with jQuery Ajax:

# Leaflet JS
var layer = L.geoJson();
$.getJSON("{% url 'data' %}", function (data) {

Inherit generic views only if you need a reusable set of options :

from djgeojson.views import GeoJSONLayerView

class MapLayer(GeoJSONLayerView):
    # Options
    precision = 4   # float
    simplify = 0.5  # generalization

from .views import MapLayer, MeetingLayer
url(r'^mushrooms.geojson$', MapLayer.as_view(model=MushroomSpot, properties=('name',)), name='mushrooms')

Most common use-cases of reusable options are: low-fi precision, common list of fields between several views, etc.

Options are :

  • properties : list of properties names, or dict for mapping field names and properties
  • simplify : generalization of geometries (See simplify())
  • precision : number of digit after comma
  • geometry_field : name of geometry field (default: geom)
  • srid : projection (default: 4326, for WGS84)
  • bbox : Allows you to set your own bounding box on feature collection level
  • bbox_auto : True/False (default false). Will automatically generate a bounding box on a per feature level.
  • use_natural_keys : serialize natural keys instead of primary keys (default: False)

Tiled GeoJSON layer view

Vectorial tiles can help display a great number of objects on the map, with reasonnable performance.

from djgeojson.views import TiledGeoJSONLayerView

    TiledGeoJSONLayerView.as_view(model=MushroomSpot), name='data'),

Consume the vector tiles using Leaflet TileLayer GeoJSON, Polymaps, OpenLayers 3 or d3.js for example.

Options are :

  • trim_to_boundary : if True geometries are trimmed to the tile boundary
  • simplifications : a dict of simplification values by zoom level

GeoJSON template filter

Mainly useful to dump features in HTML output and bypass AJAX call :

// Leaflet JS
L.geoJson({{ object_list|geojsonfeature|safe}}).addTo(map);

Will work either for a model, a geometry field or a queryset.

{% load geojson_tags %}

var feature = {{ object|geojsonfeature|safe }};

var geom = {{ object.geom|geojsonfeature|safe }};

var collection = {{ object_list|geojsonfeature|safe }};

Properties and custom geometry field name can be provided.

{{ object|geojsonfeature:"name,age" }}
{{ object|geojsonfeature:"name,age:the_geom" }}
{{ object|geojsonfeature:":geofield" }}

Model and forms fields

GeoJSON fields are based on Brad Jasper's JSONField. See INSTALL to install extra dependencies.

They are useful to avoid usual GIS stacks (GEOS, GDAL, PostGIS...) for very simple use-cases (no spatial operation yet).

from djgeojson.fields import PointField

class Address(models.Model):
    geom = PointField()

address = Address()
address.geom = {'type': 'Point', 'coordinates': [0, 0]}

Form widgets are rendered with Leaflet maps automatically if django-leaflet is available.

All geometry types are supported and respectively validated : GeometryField, PointField, MultiPointField, LineStringField, MultiLineStringField, PolygonField, MultiPolygonField, GeometryCollectionField.

Low-level serializer

from djgeojson.serializers import Serializer as GeoJSONSerializer

GeoJSONSerializer().serialize(Restaurants.objects.all(), use_natural_keys=True, with_modelname=False)

Low-level deserializer

from djgeojson.serializers import Serializer as GeoJSONSerializer

GeoJSONSerializer().deserialize('geojson', my_geojson)

You can optionally specify the model name directly in the parameters:

GeoJSONSerializer().deserialize('geojson', my_geojson, model_name=my_model_name)

Dump GIS models, or fixtures

Register the serializer in your project :


    'geojson' : 'djgeojson.serializers'

Command-line dumpdata can export files, viewable in GIS software like QGis :

python dumpdata --format=geojson yourapp.Model > export.geojson

Works with loaddata as well, which can now import GeoJSON files.


Version 1.X:

  • Daniel Sokolowski, serializer snippet
  • ozzmo, python 2.6 compatibility



  • Lesser GNU Public License