humanitarian.info

Previously. Now: there’s a new mailing list [http://lists.openstreetmap.org/pipermail/talk-ps: register at talk-ps@openstreetmap.org] for people who want to get involved in mapping in Gaza and the West Bank. They’re still looking for people with direct knowledge of Gaza to join in the editing process. A particular note for me:

In the process, agencies like UNOSAT and EC JRC have started selective release of their data sets .. a great start to open geodata exchange. The OpenStreetMap Wiki has an extensive (perhaps the greatest) collection of Palestinian geodata on the internet, all collected rapidly through crowdsourcing. [My emphasis.]

See? Crowdsourcing can work.¹ More can be found on Mikel’s blog at http://brainoff.com/weblog/index.php?s=gaza+openstreetmap.

1. Sometimes. [↩]

Following my thoughts about being mapless in Gaza, I wanted to follow up on the work of UNITAR-UNOSAT, who have made the leap from the more basic satellite images that they used to provide, and are now regularly providing damage assessments. Their analysis of postwar damage in Georgia was very interesting¹ and now they’re producing similar damage assessments over Gaza, with a commitment to update as often as they get new images.

I mentioned the .kmz file that Stefan at Ogle Earth has been putting together, which includes the UNOSAT layer. Stefan also lamented the fact that – while they provide frequent updates and quality outputs – UNOSAT products are only provided in PDF format.

And yet, the result, always, is a PDF map, which is great for printing out but not any good for any other kind of use. In some cases, the PDFs are locked against everything but printing, which means taking screenshots in order to rasterize them for placement in Google Earth… Given the global scope of these maps, their timeliness and usefulness, wouldn’t it be great if these were automatically published as KML to the Global Awareness default layer in Google Earth? People wouldn’t even need to go look for maps when they zoom in on a region hit by an emergency.

Well, I’ll agree with Stefan up to a point. PDF files are useful for nothing except printing – but most of UNOSAT’s potential users only want to print them , and playing around with the data is the last thing on their minds. However the good news is that it looks like they’re already starting – the damage assesssment data is also available as a geodatabase file and as a .kmz file. Einar has been circulating these versions to people working on the response, but has reservations on two grounds.

1. The first is regarding the added value of releasing the data more widely – what is it, exactly? My response is that to fulfill their mission as effectively as possible, UNOSAT should be producing multiple formats and distributing across various distribution channels – and a side effect of this will be an increase the possibility of useful and interesting applications emerging. We can’t predict what they might be – and they might not even appear – but the whole neogeography field is based around innovation – it just needs the data to enable it.
2. The second concern is more difficult to address – the question of whether the data will be misinterpreted or misused. This data will never be 100% accurate, which can lead to criticism of the agency publishing it if people don’t understand that. There’s also a slim chance that the data might be abused – for example, to misrepresent the situation on the ground – although the chances of this seem very small. My response to these problems is that people are free to criticise on the basis of the PDF file already, and releasing the data is unlikely to increase the type or frequency of criticism. We faced this all the time in the Humanitarian Information Centres – people would come in waving a printout and saying “Your maps are wrong!”, to which charge we would patiently explain that all maps are wrong, and would they like to help us improve?

To some extent Open Street Map have already started to deal with these issues using their existing community mechanisms, but UNOSAT is different – it’s a formal organisation in a large bureaucracy without the mandate or means to deal with public enquiries like this. Perhaps the best approach would be a tag-team of UNOSAT and OSM – sharing data as widely as possible, with UNOSAT the corporate source and OSM the buffer to address these issues as they arise?

1. Although under-utilized on the ground – that’s the next obstacle we have to overcome, guys! [↩]

Stefan delivers the Google Earth goods:

UNOSAT has just released a map dated January 9 that contains satellite imagery of Gaza City acquired by the WorldView-1 Satellite on January 6… I’ve added that map to the Gaza maps network link for Google Earth, which in the meantime also contains the updated OCHA Gaza situation map, dated January 8.

While Jon gets irate about the state of humanitarian mapping in 2009:

Impressive work all around but I need to gripe about what I see as an antiquated way of approaching humanitarian disasters at least as far as mapping is concerned… We can’t keep doing this.  We need to evolve.  There are too many people relying on us.  It is time to work past licensing issues, or whatever the real issues are, and start making substantive changes.

And Mikel defends the indefensible prospect of better maps of Gaza:

There’s a again an presumption of insider knowledge here, that anyone who is operating in Gaza is going to know what’s up. I don’t believe that… There are better things for you to do for Gazans. Don’t do this. Most of us can’t do anything directly. Actually no one from anywhere can get into Gaza to help. Why discourage a contribution?

I’m with all of these guys, all of whom are doing sterling work in trying to push the boulder of humanitarian mapping up a particularly steep hill called “business as usual”. Nigel Woof at MapAction recently asked me for feedback on the main lines of progress in humanitarian GIS in the last year¹ but I still feel like my 2006 essay on is largely still accurate – although I will agree that it fails to cover neogeography particularly well²

I still feel that our biggest problem is our lack of a clear objective in improving spatial data provision. I agree with Mikel that the OSM approach can really improve data quality in real time – but if nobody will use that data, then it calls into question the whole endeavour³ – but I would still encourage people to contribute to OSM on the basis that it is a long-term investment in a public good.

But still nobody answers my basic questions - who are these maps for, when are these maps for, what are these maps for? Maybe different projects answer different needs – but then we run into the interoperability question (I agree with Jon here – PDF files? In 2009?). Crowdsourcing alone isn’t the answer, but in this case it’s a better start than business as usual.

1. And Nigel – apologies for not getting back to you sooner. [↩]
2. Time for an update, possibly. [↩]
3. See also Ushahidi for another application of this objection. [↩]