Wazimap-ZA

The South African instance of Wazimap, a Django application for exploring census and other similar data.

Status

Note: This project is no longer being supported by OpenUp. We are building a more modern and richer tool which we recommend new users choose instead. You can find it here:

• https://github.com/openupsa/wazimap-ng
• https://github.com/openupsa/wazimap-ng-ui

Local development

1. clone the repo
2. cd wazimap_za
3. virtualenv --no-site-packages env
4. source env/bin/activate
5. pip install -r requirements.txt

Set the WAZI_PROFILE environment variable to the instance you are working on, e.g. export WAZI_PROFILE=census

Set the DEFAULT_GEO_VERSION environment variable if you don't want to default to the latest, e.g. for youth export DEFAULT_GEO_VERSION=2011

You will need a Postgres database for the instance you are running:

psql
create user wazimap_<instance_name> with password wazimap_<instance_name>;
create database wazimap_<instance_name>;
grant all privileges on database wazimap_<instance_name> to wazimap_<instance_name>;

Import the data into the new database.

cat sql/*.sql | psql -U wazimap_<instance_name> -W wazimap_<instance_name>

Run migrations to keep Django happy:

python manage.py migrate

Import the fixtures for the django models:

python manage.py loaddata fixtures/<instance_name>/wazimap_django_models.json

Create an admin user:

python manage.py createsuperuser

Start the server:

python manage.py runserver

Updating django model data:

Be sure to dump the data to the appropriate fixture when making changes to django models data:

python manage.py dumpdata wazimap > fixtures/<instance_name>/wazimap_django_models.json

Production deployment

See the Wazimap deployment docs for all basic Wazimap configuration.

Set the profile

dokku config:set wazimap_za WAZI_PROFILE=<profile name>
dokku config:set wazimap_za DEFAULT_GEO_VERSION=<geography version>

Add dokku as a remote, and then deploy:

git push dokku

Adding new census data

Use the python manage.py importcsv command to import CSV data exported from StatsSA using the SuperWEB or SuperCROSS packages.

Let's suppose you want to add a new table with two fields: favourite colour and gender. We'll describe the process for SuperWEB, but SuperCROSS is similar.

Downloading from SuperWEB

1. Visit SuperWEB
2. Open the data collection that contains your data (Generally under Community Profiles > Census 2011) by double-clicking the database.
3. Add the geography details as rows by clicking the Geography heading and choosing the levels you want to import. To keep the amount of data sane, export wards separately to other levels such as municipality and province.
4. Check the boxes alongside the levels you're interested in (or click the arrow to the right of Geography and choose the correct child level)
5. Click the Add to Row button
6. Find the field for the first field you're interested in (gender, in our example) and ensure its children are selected.
7. Click the Add to Column button.
8. Do the same for the second field (if appropriate): favourite colour in our example.
9. Note: when adding more than one column, be sure to turn off automatic totals for columns by clicking the little blue Sigma symbol to the right of the field name (eg. Gender) in the column.
10. Click the big play button Retrieve data to load data into the table.
11. Download the table as Comma Separate Value (.csv) (not CSV string value).

Importing into Wazimap

1. Ensure wazimap_za/tables.py has a FieldTable that has exactly the columns that you're importing. If there are multiple tables with exactly the same columns, perhaps because their Universes are different, then be sure to take note of the table id.

2. Do a dry-run of the import, using the table name if necessary.

    python manage.py importcsv yourfile.csv --dry-run [--table TABLENAME]
   

3. If it all looks good, run it without --dry-run.

4. Update (or create) the raw SQL data:

    python manage.py dumppsql --table TABLENAME > sql/TABLENAME.sql
   

5. Commit to git.

6. All done!

To dump all data tables at once, run

for t in `ls sql/[a-z]*.sql`
do
    pg_dump "postgres://wazimap_za@localhost/wazimap_za" \
        -O -c --if-exists -t $(basename $t .sql) \
      | egrep -v "(idle_in_transaction_session_timeout|row_security)" \
      > sql/$(basename $t .sql).sql
done

License

MIT License