Slightly bad data vs. no data at all

At the BarCamp Africa UK, a good many of us had a common problem in that while the event took place at at the Vodafone headquarters in a very slick, new building, Google Maps was nearly no help in finding it. The actual address is 1 Kingdom St,London W2 6BL. As you can see in that link, it does indeed pull up an address that is near the event, but is not the actual event. That address points you to a series of row houses on the other side of the train tracks going to Paddington Station. The actual location, I’ve shown below. As you can see, it’s close, but no cigar.
The quick, obvious moral of this is that you can’t entirely trust Google Maps. But, that is getting harder not to do as we rely on technologies like this for a great deal of our information because more often than not, they are a great deal more correct than traditional sources. While Google is taking a great deal of effort in trying have more community involvement in their mapping projects, you still end up with maps like this one of central Mombasa. Sure, whatever automatic system they used to generate this worked decently well, but at the same time, in the map version, there are streets going across the water! And of course, a lot of things are unnamed. While many would say that this is a cost of crowdsourcing which is far outweighed by the eventual benefit of more data, unlike Wikipedia, Google is a corporate entity. They make money off these maps (through related ads) and as such, I believe that there is a good deal of responsibility to make sure that things are accurate. For anyone who is in development communities, you know very well the difference between the “nightly build” which is often bleeding edge and very much broken vs. the “stable release”. Google seems to be releasing their “nightly build” for general consumption. Yes, I realize that it may encourage others to come and fix it, but what about those who see it, see that it’s Google and take it as cartographic gospel?
While Google is one of the few companies really striving to have their products work in Africa, it seems that there is an underlying principle of “good enough” in some of what they’re deploying. Because the fact that there were no digital maps of Mombasa or other cities available before, what they’re offering is better than that because it’s “something”. But all this is doing is quashing some kind of local group or even OpenStreetmap (their view of Mombasa) to flourish.
Google and other companies need to really be called out on things like this and have a great deal more accountability. If their maps other other data are going to become the de facto, then someone needs to certify that things actually are correct because in my book, you simply cannot half-ass data collection; it’s full-ass in or don’t do it at all.
Slightly bad data vs. no data at all

One Reply to “Slightly bad data vs. no data at all”

  1. Having worked in the visa building for two years and watched the Vodafone offices being built I know the area really well. This is really a case of google maps not being updated for the new roads – though I imagine it’s also a case of the Royal Mail not updating their postcode and geo-location information as Google relies on that somewhat for the actual locations.

Comments are closed.