Mildura is a town in the far west of the state of Victoria, Australia. According to Google/Wikipedia, around 30,000 people live there. News today, however, that for the rest of the world, Victorian police are warning those looking for Mildura not to use Apple Maps.
Mildura Police are urging motorists to be careful when relying on the mapping system on the Apple i-phones operating on the iOS 6 system after a number of motorists were directed off the beaten track in recent weeks.
Local Police have been called to assist distressed motorists who have become stranded within the Murray-Sunset National Park after following directions on their Apple i-phone.
Tests on the mapping system by police confirm the mapping systems lists Mildura in the middle of the Murray Sunset National Park, approximately 70km away from the actual location of Mildura.
Police are extremely concerned as there is no water supply within the Park and temperatures can reach as high as 46 degrees, making this a potentially life threatening issue.
And if you try it out, here is what you get. Mildura is for some reason deep inside a national park.
Go to Google and they have it right.
So how to Apple lose a town? The current view is that the bulk of Apple’s Map woes are coming from poor data and the need to have use that allows users to identify inaccuracies. Google have years on Apple and so the concern here is that Apple will need those years to catch up.
But the Mildura issue flies in the face of that. If you want to report to Apple where Mildura is, you can from their maps app but it prominently tells you that its data comes from TomTom among others. So could it be that TomTom don’t know where Mildura is?
No, here is what TomTom show you:
Mildura is where it should be. In other words, Apple must have the right data. What it has failed to do is implement it properly with its search tools within maps. But one has to believe that that is a solvable problem and indeed one that can be solved fairly quickly. It also demonstrates precisely why Apple moved to revamp its Maps leadership team. It is a problem that should have been solved earlier. Nonetheless, the point here is that, if this is the issue with the bulk of Apple Maps problems, then Apple will end up solving it sooner rather than later. Now it may be that isn’t enough but TomTom have been in this game for much longer than Google. It seems to me that they surely have the ability to provide an accurate map. It just has to be programmed correctly.
Oh yes, and those 30,000 people. When I asked Siri the same question, here is was the answer:
This answer (from WolframAlpha) is closer to the Mildura’s local council estimates. The 30,000? That’s from the 2006 Australian Census. Both Google and Wikipedia are 6 years and 66% out of date. It goes to show that it is far from clear that our traditional incumbent search engines are best of breed.
Digitopoly 
Its not clear to me that the 30,000 estimate is incorrect when the link you provide also lists that as a correct number. Estimated resident population vs Mildura Population. The answer to “what is the population?” can lead to estimates or census figures. It all depends on what you are looking for. The location of an entire city is typically less debatable. Wolfram and Siri don’t report the possible range given that it is an estimate and it is 4 years old.
Lets not forget that a city council is likely to bias towards overestimating their towns population.
I’m not sure you’ve established in what way it’s good news for Apple.
The key problem is how to make sense of multiple sources of data, and then contextualise it. I’m sure this particular example will now be corrected quickly, but I don’t agree that the root cause of the problem will necessarily be resolved “fairly quickly”.
Josh’s argument is that connecting data in different ways is likely to be a fundamentally easier task than gathering the data in the first place. It may be (though probably not) that Apple simply needs to correct one line of code in the programming that feeds data out of TomTom’s system into Apple Maps.
If Apple had to collect data on Mildura and X other towns, this argument goes, they’d have five years’ work ahead of them.
This seems to me more likely than not to be true.
I think the problem here is apple maps (and Wolfram Alpha and therefore Siri) is returning data for the local government area, whereas Google maps and Wikipedia returns data for the regional city. The pin in the Apple map appears to be in the middle of the local government area. See:
http://en.wikipedia.org/wiki/Rural_City_of_Mildura
vs
http://en.wikipedia.org/wiki/Mildura
It also explains the discrepancy in population size:
http://www.censusdata.abs.gov.au/census_services/getproduct/census/2011/quickstat/LGA24780?opendocument&navpos=220
vs
http://www.censusdata.abs.gov.au/census_services/getproduct/census/2011/quickstat/SSC20893?opendocument&navpos=220
I think it’s obvious in this particular case that searching for Mildura you want the city rather than the area, but the general case is much, much harder (and it’s what Google is really good at).
(2011 census population)
Mildura: 31k
Irymple (almost a suburb of Mildura): 6300
Red Cliffs (a district adjoining Irymple): 5600
Merbein (district adjoining Mildura): 4500
This accounts for the discrepancy. All these districts are within commuting distance of each other.