humanitarian.info

Nigel Woof at MapAction has just circulated a briefing paper entitled Google Earth and its potential in the humanitarian sector [pdf], which outlines most of the key issues around the use of GE (and other geospatial visualisation tools). I was particularly pleased that Nigel recognises the way in which GE is a disruptive technology, something which I discussed a couple of years ago in Here comes the Geographic Information Revolution (in fact I made a bet on it, but needless to say nobody noticed).

There were a couple of points which I thought were worth picking up on. Nigel notes that it is difficult to work with polygon attributes in GE, despite the fact that these are frequently the most useful basic patterning tool. This is absolutely right, and until Google addresses this issue, it’s going to be the single largest obstacle to its effective use for operations. I ran straight into the problem in Bangladesh – built the polygons for administrative boundaries from a shapefile, but adding the attributes that I wanted to was just more effort than it was worth, so I stuck with the Scrappy Maps.

Pull quote:

When producing anything other than a very simple map, Google comes nowhere near GIS for sheer efficiency and flexibility of cartography. But the huge advantage of using Google Earth is that non-experts can use it in an intuitive way to visualise relatively simple data, without having to worry about georeferencing or editing complex geometry.

Absolutely, and we should all welcome the massive benefits that GE brings with this simplicity. However I am also fairly certain that these benefits do have a cost – and that cost is properly managed geospatial data. Anybody who’s worked with spatial data knows how much work goes into maintaining it, but GE creates a disincentive to manage data in a systematic manner.

While this is fine in the short term, it doesn’t bode well for the creation of spatial data infrastructure that can be put at the service of a wider range of users. This is an ongoing discussion that won’t be resolved soon – discussions of the type that OpenStreetMap are having right now. A completely distributed approach to geospatial data won’t work, but neither will a wholly centralised approach; we need to navigate somewhere inbetween.

There seem at present to be two distinct groups of humanitarian practitioners: those who are already, albeit tentatively, exploiting Google Earth and related geospatial methods in their work, and those who will be, as soon as they see their first demonstration of its potential.

I am a little less optimistic – I think there’s a third group that will reject all attempts to introduce GE as a tool into their work, based on the experience of trying to get people to participate in GIS development at even the most basic level. There are a large number of people who want maps but have no interest in the process itself. We can’t assume that everybody will jump on the GE bandwagon, beyond the initial “wow, that’s cool” moment when they first see it.

So where are we, and where do we go from here? Nigel’s paper is not very specific, beyond recommending that we invest in GE as a tool as the basis for more user-oriented mapping activities. He’s absolutely right of course, but saying that GE has little to no cost is not true – there are plenty of hidden costs such as the one described above. We should welcome the potential that GE has for turning everybody into a mapmaker – but we should remember that different people’s maps don’t always agree with each other.

[UPDATE: Christiaan Adams has posted Nigel's paper at the Google.org blog, which is great - more visibility for this sort of analysis in the wider world!]