[HOT] Meeting with MSF UK

Harry Wood mail at harrywood.co.uk
Fri Dec 7 18:02:23 GMT 2012

Last week I met with people from MSF UK ( http://www.msf.org.uk  Medicins Sans Frontiers - Doctors Without Borders )  This is to relay some of points discussed.    TL;DR: there's lots of ways MSF and HOT can work together. Mapping DRC is a priority right now.

The London office supports emergency missions through it's specialist medical unit (Manson Unit). The response coordination desk is elsewhere, but this office supports both short term and long term emergencies and has strong role in supporting operational research in MSF, meaning they try to come up with innovative solutions to field problems. This includes a small team developing some GIS capabilities.

We spent some time talking about DRC. MSF are active here right now. Insecurity and violence in eastern congo continues, and clearly there are lots of humanitarian efforts ongoing there. It's a high priority area for mapping. This job http://tasks.hotosm.org/job/102 (which mostly looks finished actually), and other mapping progress in DRC could be useful to MSF.

I told them about aerial imagery being a limiting factor for our remote mapping, and the twin challenges: permission to derive vector data, and technical (geo-rectifying, tiling and hosting)  They may be able to help formulate requests for imagery in important areas. We discussed the potential of aerial photography techniques such as balloon photography, so MSF may eventually own some of their own imagery resources. I imagine the technical aspects of that are an area we can help with.

Actually while discussing I had imagined that we had approximately zero bing imagery available in DRC, but I see there *are* some strips on there now: http://bit.ly/DRCbing  Would be good to get a clearer picture of that.

We talked about various GIS use cases, of which only some are playing to the strengths of OpenStreetMap. A lot of data they could really benefit from, is the kind of data which we haven't been able to get hold of either. They'd like to do fairly zoomed out choropleth maps, for which the main data they'd need is some administrative boundaries. They also need maps which show where villages are, with names. We talked about how GNIS data is sadly lacking precision, leading to weirdness like thishttp://osm.org/go/wE9IcVt--  (interestingly this area of Nigeria has new hi-res bing imagery since I last looked at it)  I tried to stress that if they do find sources of improved data e.g. their own field workers, then getting this into OpenStreetmap is a very effective way of sharing it.

Around all of these topics I described OpenStreetMap in detail. I also demonstrated JOSM, and talked about offline capabiities. I also described some of the aims HOT, to get out on the ground ourselves as part of humanitarian responses.

There are areas where we can certainly help them with data or with data gathering processes. They have done work on population estimates by counting buildings, so our Indonesia disaster preparedness work serves an example of doing this the OSM way. We can work towards serving their use cases with remote mapping, even where we can't go the whole way e.g. village maps without names, and ensuring basic natural features (coastline and large rivers) are mapped correctly from low-res imagery for the purposes of basic medium-zoom basemaps.

Of course I invited her to bring us more specific mapping targets when they have them, and she has shared a list of the mapping priorities they have "where we eventually need our operational areas base-mapped to the village level" 

The other thing shared with me was a link to this joekit system they've been building. A java program to go from spreadsheet data to KML. 
I imagine you all could bombard them with a million other suggestions for ways of solving this problem better/differently. But at least we can be heartened to see that they're not afraid to solve a problem with a bit of coding. We can definitely teach these people to become OSMers :-) Also this gives a feel for at least one class of GIS problem they're tackling: Very zoomed out maps of disease outbreaks.

Clearly there's lots of ways MSF and HOT can work together. We've resolved to stay in touch and meet again soon, so I suppose if people have suggestions for further things I should raise with them, let me have them.

Harry Wood
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20121207/ce911e63/attachment.html>

More information about the HOT mailing list