Data is Chaos

One of the greatest things about the Internet is the vast amounts of data available on almost every aspect of our existence on this planet. The other beauty is the journey that is taken searching for the data, starting with spirals that attempt to reach their tails until eventually falling into the cracks that manifest into endless fractals of potential, and all the while allowing for the searching of patterns within all the chaos. In a strange sense (or at least strange to most people), that is what this site is about, and so we think it's important to point to the places where data can be discovered.

Website

Firstly, the site was built using GeoDjango which is a brilliant geo contrib-library for the Django web framework. We use jQuery as our javascript library, Postgres with PostGIS for the database, OpenLayers with Google Maps to display maps, and Apache for the front-end.

The site was lovingly hand-coded using TextMate, and the various graphics where made using Adobe applications. For the icons we didn't design, we mostly used the Nuvola icon theme.

Data Sources

As we had started the site with the intention of using the 2006 Australian Census data, we figured it would be best to start from there. As a result, the initial 9,152 towns/regions came from the ABS, which includes the place names, census data, and general census codes. We have also used Shape files available from the ABS, to set the initial coordinates for each location (using the centroid for the shape polygon), matched town polygons to postcodes, and then finally matched towns to regions by analysing polygon intersection percentages. Coordinate data is stored as EPSG 3112 (GDA94/Geoscience Australia Lambert) to accurately measure distances and areas within Australia.

We next used GeoNames to fill out the remaining towns not included in the ABS data. From the GeoNames database, we also used their coordinates to check our initial ABS block of data, and through a very gradual (and painful) process, managed to replace the ABS coordinates with the data from GeoNames. For the coordinates that were a little more troublesome, we first programatically used the Google Maps API to find and set coordinates, and then started to set them manually by double-checking the Geoscience Australia Place Name search. Postcodes were checked and later double-checked against the GeoNames database, Australia Post, various Government websites, as well as PDF files available on the Medicare website.

Town details were found using so many sources, we honestly can't remember where everything came from. We did reference sites like Bureau of Meteorology, Wikipedia, National Geospatial-Intelligence Agency, etc.