Justin O’Beirne has long believed that the future of map making has to be intertwined with technology, and his career has seen him work in Cupertino contributing to Apple's Maps platform. He's also an avid blogger and public speaker — he did a talk at Google I/O back in 2011 about styling digital maps for enhanced usability.
On his website, O'Beirne likes to track the progress of different map apps, comparing them against previous versions and each other. His most recent post looks in depth at Google Maps and just how far ahead of the rest of the pack it is. I think most of us perceive that to be the case – even putting aside our obvious Google bias – but without being able to say exactly why. That's where O'Beirne comes in.
To explain what Google is doing to be better than other digital mappers, O'Beirne first considers the amount of detail each service goes into. As Google Maps users we probably take for granted just how true to life it can be, and this is blatantly apparent when you take a closer look at buildings. Google started adding buildings as far back as 2007, and it's at a point now where even small towns with a population of less than 1000 have them, as well as other smaller structures like sheds and park shelters.
In stark contrast, Apple still doesn't have buildings in many major cities, nor does it correctly distinguish green spaces. It's not just Apple that has fallen behind in this respect, though, with Bing, Here, TomTom, Mapbox, and others all struggling to offer the same level of detail as Google.
Even when competitors do have buildings, they're nowhere near as detailed as Google's. You can often see the exact shape of a structure on Google Maps, at least in big cities, complete with design elements and features such as air conditioning units and satellites – a far cry from Apple's rudimentary outlines. The Google Maps team first started this process by adding around 1000 accurately rendered landmarks back in 2012.
The same algorithmic computer vision techniques outlined in that Google post are now being put to use across all building types. It's amazing to think that Google is simply utilizing the same aerial imagery it has already captured to do this. That Google has added buildings with such great detail to US towns that haven't even been fully explored by Street View cars yet is an incredible feat, and one that ensures Google Maps is many miles ahead of any competitor.
The blog post goes on to praise Google for its use of special shading to highlight "Areas of Interest" (AOIs) at different zoom levels, avoiding the overuse of labels that get in the way. Using a combination of building and place data, Google is able to generate these AOIs algorithmically, and with stunning accuracy. Add in Street View data, and we see that Google Maps is a ridiculous achievement that is only possible by understanding and combining vast quantities of different datasets, which clearly only Google is capable of at this level. It's the only company that has all of the ingredients as well as the know-how to put them all together.
O'Beirne goes even deeper into the challenges Apple and the others face in closing the map gap. He estimates that Google might be as many as 6 years ahead of Apple in AOI data collection, thanks to its "Ground Truth" extraction project that commenced in 2008. He also goes on to speculate about other advancements Google could be using the data for, such as a self-driving ride-hailing service. You should head over to the post to read more. By making data out of data, Google has given itself such a big lead in digital cartography that it's hard to see how anyone could catch up with it. I'm going to try to appreciate Google Maps a little more when I use it from now on. I've certainly been taking it for granted.