Wednesday, 23 May 2007

Another Reality Zap

One of the goals for SDI-East Africa has been to improve service between the UN bodies in this part of the world and their "external" (a.k.a. non-UN) partners. One of the early potentials has been the Nairobi office of the International Committee of the Red Cross (ICRC). Not only are they major user of geo-spatial data but their extensive field operations mean that they need to share and coordinate on common needs with the likes of UNHCR and OCHA.

So, back on March 26, I noted that we were to be heading off to build some capacity at ICRC. This we then did and, so impressive were the results that we were invited back on Monday to get Geoserver running on a publicly-visible web site. And this is where the proverbial hits the fan, not over technical issues but institutional ones.

ICRC are very rightly concerned about the integrity of their global network operations. These are coordinated from Geneva and are subject to policies originating there. The network administrator's here in Nairobi operate and manage a "black box" that proxies for users on the unsecured portion of their LAN, but does not provide any means for announcing a web service per se. (ICRC Nairobi do not have their own web site, for example). So although Java, Tomcat, Geoserver and the PostgreSQL spatial database back-end could all be installed and demonstrated to be running, there's no way that an outside user could address that machine and see the services.

This, for the first time in the course of the SDI-EA exercise, exposes and instance of where an institution has a programmatic imperative to publish data and is prepared to establish and maintain the services to make this possible, but runs afoul of higher-level institutional policies. I doubt that this will be the last such instance. For ICRC Nairobi it raises two potential courses of action, one being to get ICRC Geneva to review and revise their network policy, the other being to see an outside host for their outward-looking services. Most fundamentally, though, it demonstrates how moving into a web services framework can (and probably will again) expose tensions between different aspects of institutional philosophy. It will be interesting to see how ICRC choose to resolve these.

Monday, 7 May 2007

A Satisfied Customer!

I know this is likely to come across of blowing one's own trumpet but, on the other hand, when there's a happy and satisfied customer who sees further potential and value flowing from their first foray into this interoperability stuff then that's an important message to get out. Part of the hope of SDI-EA since the outset has been the idea that being able to take incremental steps and demonstrate success every step of the way is key to ensuring management support and community acceptance for these SDI-type activities. Maybe we start seeing justification for that approach.

Anyway, the short version of the story is that UN High Commission for Refugees (UNHCR) prepare periodic reports based on field asessments of the movements of internlly displaced persons (IDPs) in Somalia. As you know from the news hundreds of thousands have fled Mogadishu in recent months and the humanitarian agencies and NGO's scramble (sometime literally) to keep their supply lines up to those in need. UNHCR have distributed the location of the IDP camps as lat/lon coordinate pairs coded in Excel sheets and/or as accompanying maps.

What DEPHA (the Data Exchange Platform for the Horn of Africa, offered to do with some support from UNEP was take those data and publish them as an open-standard Web Feature Service. One consequence? The same data can be output as KML and viewed via Google Earth. And our man at UNHCR is very happy about this, and sees real utility in not only having a new vehicle for communication but also for integration with other types of data. Download or try the WMS at, or try the WFS at

As an aside, a colleague here at UNEP who saw UNHCR's data coming up last Friday was very quick to see the potential for integrating with his existing ArcIMS service depicting refugee camps across Africa. He was even more impressed when he saw that he didn't need to copy and integrate the data but could just link them in via the WFS interface. Yea. Interoperability. Getting our own back yard in order.

Now all we have to do is get the IDP data showing up in the context of UN Habitat's urban areas data and I will be glad to say that this humble little SDI-EA effort will have passed a major milestone.

PS FAO, UNEP and DEPHA are contemplating a (couple of?) hand-on session(s) for technical types where we'd actually set up a network and have practitioners bring their data, set up their own WFS's and start interoperating at least on a local scale. As usual there's no budget or blessing for this but we figure that we can scrape together enough PCs and network gear and a meeting room for half a day to make it work. The whole session would not take more than a morning, starting from scratch with the server intllations (takes about 10 minutes), setting up the spatial databases, and transforming existing static data into exciting dynamic content being integrated across institutional boundaries. Think of it as a "train-the-trainers" exercise; if we're successful then, who know, maybe RCMRD might offer to host a followup at Kasarani, or add such offerings to their schedule of courses.


Mick Wilson
Division of Early Warning and Assessment (DEWA)
United Nations Environment Programme
PO Box 30552 - Nairobi 00100, Kenya
Tel: +254 20 7623436
Fax: +254 20 7624315

----- Forwarded by Mick Wilson/UNEP/NBO/UNO on 05/07/2007 04:19 PM -----
"John Marinos"
05/07/2007 03:12 PM

, ,
Re: Fw: IDP settlement coordinates to be hosted by DEPHA..also onGoogleearth

Thanks John,

I think this looks great. Really. I will share with our HQ because they love stuff like this.
It seems I need to get an update to my ArcGIS before using the WFS files. Nevertheless, I think the information on Google Earth is super-cool.

In theory there is no reason we couldn't merge this GIS file with a whole bunch of attribute data - pop, major area of origin, whatever... I suppose that this can easily be updated on GE, right? We'll work on that when I'm back from leave.

Thanks again. Seriously this is really good.


>>> John Mugwe 05/07/07 02:40PM >>>
Sorry all for that UNEP logo...It was just a proof of capabilities.

We have now a new kml file ( see attached ) which is also going to be
uploaded on by close of the day

(See attached file: idpSomalia.kml) where anyone access it.

Best regards

John Mugwe