I have visualised each postcode as a scatter plot using Plotly / Mapbox and this shows immediately that a few localities have dodgy data in the Google / precise coordinates columns. See the HTML plot in the zip file attached.
The dodgy ones I can see are:
Tarrawarra QLD (coordinates are for Tarrawarra VIC) but should be a suburb of Cairns
Diamond Head QLD coordinates place it firmly in NSW
Martins Creek, Minto, and Yarrunga (all VIC) have coordinates placing them in NSW
Coordinates for postcode 0872 ("ANANGU PITJANTJATJARA YANKUNYTJATJARA") also appear to be firmly in NT whereas the data describes it as being in SA
Berrambool VIC mixes up its "precise coordinates" and data with Berrambool NSW
Deakin University 3217 (record 6684): it seems this is referring to the campus in Geelong (Waurn Ponds) but the SA and Google Maps data has found the Burwood campus.
Happy Valley QLD 4655: seems this is a minor area sharing Hervey Bay's postcode but data was for Happy Valley 4825 near Mount Isa
Happy Valley 2430 has coordinates in SA
Christies VIC 3920 has coordinates in SA
In all of these cases the places don't appear in Australia Post's postcode database or Google Maps but they do appear on http://postcodes-australia.com. (They seem to be minor suburbs that share a postcode with others.)
Not sure of your workflow for contributions. I'd be happy to fix them in the CSV and make a pull request but the repo has so many file formats it would be far more work to fix them all. I assume you have some automated process to generate the different file formats from one canonical source?
Thanks for this awesome repo! Very useful.
I have visualised each postcode as a scatter plot using Plotly / Mapbox and this shows immediately that a few localities have dodgy data in the Google / precise coordinates columns. See the HTML plot in the zip file attached.
The dodgy ones I can see are:
In all of these cases the places don't appear in Australia Post's postcode database or Google Maps but they do appear on http://postcodes-australia.com. (They seem to be minor suburbs that share a postcode with others.)
Not sure of your workflow for contributions. I'd be happy to fix them in the CSV and make a pull request but the repo has so many file formats it would be far more work to fix them all. I assume you have some automated process to generate the different file formats from one canonical source?
visualisation.zip