Putting Django on the Map

GeoDjango

• Gareth Lloyd
• @godawful
• glloyd@gmail.com

“Everything is related to everything else, but near things are more related than distant things.” - Tobler’s first law of geography

• GeoDjango gives a nice interface to 2 C++ libs, GDAL and GEOS
• GDAL is used for getting data in and out
• GEOS manipuates spatial data
• KML/GeoJSON
• have model fields for Point and Polygon

Example app - 3sqaure

• install PostGIS http://postgis.net/install , install libs
  • MySQL and SpatiaLite have some support, but much less than Postgres
• add django.contrib.gis to INSTALLED_APPS

from django.contrib.gis.db import models

class xyz(models.Model):
    # ...
    location = models.PointField()

    objects = models.GeoManager

Create objects with shapes and points. Then can do stuff with querysets like:

Place.objects.filter(locations__within=london.shape).count()
gps_reading = Point(...)
Place.objects.all().distance(gps_reading).order_by('distance')
Place.objects.all().distance(gps_reading).order_by('distance')[0].geojson()

Many operations on shape - spacial relationships stuff.

Play with open datasets:

• ordinance survey
• data.gov.uk
• https://www.sharegeo.ac.uk/
• http://openstreetmap.org
• https://github.com/gareth-lloyd/postcode_latlng - convert PO database from weird format to usable format