This is the backend for Wazimap-NG.

Introduction

Wazimap-NG is the next version of Wazimap. It provides a platform for users to bind tabular data to spatial boundaries in order create curated views of datasets. Yes - that's probably too vague a description to understand what it is. Hopefully the images below provide a better description:

See a link to the beta site here: https://beta.youthexplorer.org.za.

You can find the backend code in this repository. The frontend is available here: https://github.com/openupsa/wazimap-ng-ui.

New features

The main new features are:

• Admins now have more flexibility when it comes to loading data. This includes uploading massive datasets and then slicing and dicing in the backend rather than pre-preparing datasets beforehand.
• Point data is now fully integrated as a first-class spatial object.
• Choropleths built into the main view. These were hidden behind multiple clicks in the previous version.
• One platform can host multiple profiles off the same database.
• The Rich data view allows richer disaggregation of indicators.
• The administrator can configure the view to use custom basemaps, colours, and other UI settings.
• Arbitrary spatial boundaries and hierarchies can be loaded onto the same server.
• Toggling of overlapping boundary layers such as switching between wards and mainplaces which typically cover the same areas.
• Integration into third-party data sources for realtime data feeds.

Related software

There is no shortage of mapping software available, both commerical and open-source. Wazimap focuses on providing a platform for data custodians to showcase their datasets and mashing them up with public data. The most similar tool that we have found is the excellent GeoNode. We feel that approach to publishing data is significantly different enough to warrant a separate project.

Roadmap

Version 0.8 is due soon and will fix bugs that currently don't have workarounds. We'll publish the 1.0 roadmap soon.

Future features

• WFS endpoint for publishing data to other GIS software
• Pluggable data visualisations
• Better handling of geography hierarchies.
• Improved handling of temporal and other types of non-census-like data.
• Speed improvements
• A large standard database of public datasets.

Prerequisites

• Docker

Local Development

Local development is normally done inside docker-compose so that the supporting services are available and the environment is very similar to how the application is run in production.

Make docker-compose start the supporting services

docker-compose run --rm web python wait_for_postgres.py

Migrate, load development data and start the app

docker-compose run --rm web python manage.py migrate
docker-compose run --rm web python manage.py loaddata demodata.json
docker-compose up

Run the tests

docker-compose run --rm -e DJANGO_Q_SYNC=true web pytest /app/tests

Run Django manage commands inside docker-compose, e.g. create a superuser:

docker-compose run --rm web python manage.py createsuperuser

Demo data

Demo data should consist of the smallest amount of data needed to demonstrate as much functionality as possible with the minimum effort. Updates to demo data should aim to keep diffs as small as possible as well.

The demo data creates a superuser with username admin. Ask for the password in the dev channel or update it using manage.py changepassword or manage.py createsuperuser.

Use the following command to generate demo data, diff to check that the change looks sane before committing. Update the command here if changes are needed.

docker-compose run --rm web python -Wi manage.py dumpdata \
  --indent 2 \
  --natural-foreign \
  profile \
  auth.group \
  auth.user \
  guardian \
  datasets \
  boundaries.geographyboundary \
  points.category \
  points.location \
  points.profilecategory \
  points.theme \
  general.metadata \
    > demodata.json

Configuration is done via environment variables.

*** Avoid creating/using variables that define an environment ***. Instead of creating a variable DEV or TEST or PROD and then writing code like if TEST, rather make each feature/backing service configurable, and configure them as needed in each environment.

Ensure variable defaults are safe if someone forgot to set them in production. Variables that absolutely must be set in each environment can be left without a default so that the app only starts when configured properly.

Key	Default	Type	Description
DATABASE_URL	undefined	String	postgresql://user:password@hostname/dbname style URL
DJANGO_DEBUG_TOOLBAR	False	Boolean	Set to True to enable the Django Debug toolbar NOT ON A PUBLIC SERVER!
DJANGO_SECRET_KEY	undefined	String	Set this to something secret and unguessable in production. The security of your cookies and other crypto stuff in django depends on it.
EMAIL_BACKEND	django.core.mail.backends.smtp.EmailBackend
DEFAULT_FILE_STORAGE	undefined	string	e.g. storages.backends.s3boto3.S3Boto3Storage or django.core.files.storage.FileSystemStorage
AWS_ACCESS_KEY_ID			Only required if DEFAULT_FILE_STORAGE is storages.backends.s3boto3.S3Boto3Storage
AWS_SECRET_ACCESS_KEY			ditto
AWS_STORAGE_BUCKET_NAME		string	ditto
AWS_S3_REGION_NAME		string	ditto
AWS_QUERYSTRING_AUTH		string	ditto
DJANGO_Q_SYNC	False		Usually useful for automated tests
ENVIRONMENT_NAME			Only to be used for a server to identify itself e.g. in logs or notifications, and not intended for configuration logic
DEBUG_CACHE	False	boolean	Set to true enable the cache despite DEBUG being true. This is useful when you want the cache enabled for development.

Documentation

These are works in progress:

• Technical manual
• Administrator manual

Contributions

Contributions are welcome - we are working towards making this process easier. New development takes place in the staging branch

Shoulders of giants

This project is the next iteration of a number of excellent projects starting with CensusReporter and Wazimap that followed it. Special thanks to William Bird from Media Monitoring Africa whose initial idea (and funding) it was to build a tool to help journalists better understand areas they were reporting on. Also thanks to Chris Berens from VPUU who directed funding to help kickstart this new build. Finally, all of the amazing spatial software and tools developed by one of the most dedicated open source communities out there.