Tuesday 8 January 2008

Life and Times in Nairobi

The post-election violence in various parts of Kenya was front-page news for at least two or three days, a remarkably long period. Kenya will be months or years recovering from the aftermath, one dimension of which is more than 200,000 people estimated by the UN to have been displaced. Setting up camps and safe-guarding the inhabitants, ensuring food and medical supplies, establishing and maintaining sanitation are task that will keep UN agencies, NGOs, IGOs, church groups and the Government engaged long after the story has ceased to be even page 10 newsworthy.

The crisis has - no surprise here - exposed a number of shortcomings in the management of the information needed to plan and manage responses. The story has been the same old confusion, duplication, crossed efforts and lost time that always seems to mark such times. Yes, of course, crises always occur at the most inconvenient time and the circumstances in Kenya were particularly confounding - key staff are out of town or out of the country because it's still holiday time here means that many ; the UN in Nairobi effectively closed down for the whole Christmas - New Year week and then stayed closed for the first week of January, with staff ostensibly 'working' from home so no access to servers or datasets that weren't visible to the internet; some key staff who happen to have data on their computers can't get to office safely. Mostly, though it exposes the fact that few agencies as yet manage and publish their data in ways that enable true use and re-use a la SDI.

Pertinent points include:
  • UN bodies in New York seeking to upgrade their security plans for Nairobi, and looking for the data used to create the last set of maps, couldn't find them and contacted me (at home) by e-mail (on a Friday night) wondering if I could find the right people they think might have had the right data;
  • the custodians of the data respond, but one's out of town on holiday, one is in town but cannot get to the office, and anyway the data are on the hard disk of a third staff member's PC and they don't have the password;
  • the data were never published to the web, nor was their metadata, because they were prepared as a small contract job and no-one thought there'd be any re-use. Whoops.
  • a colleague in another agency not only has the requested data (not just the PDFs) but even better built up for their own purposes since, but shouldn't distribute them as per agreement with their originator... but then again his agency has rectified the data and extended their attributes so, yes, okay, he sends the data to New York, who are appropriately appreciative...
  • ...except that the data dispatched are in a proprietary format because that's what they use in their office. Does New York use the same software? I have no idea but would think it a damn shame if NY couldn't use the data for lack of a software license...
  • ...and, meanwhile, a third colleague charged with coordinating information management for humanitarian response is sending out requests to the same people for the same type of data, admittedly for different application but still involving the same players in more work.
They foregoing highlights just how far we have to go with an exercise like SDI-EA before we can claim real legitimacy:
  • where is the one-stop shop catalogue discovery system that would enable to New York to at least confirm the identity of the custodian of the original data? Such catalogues are alive and kicking in Nairobi (see FAO/SWALIM's GeoNetwork node) as resources for the community - why aren't they being used?
  • where are the metadata being routinely published that describe the sort of interim data products that underlay the UN security plans, that in a very real sense have been produced using taxpayers money, but are allowed to languish in dark cupboards?
  • where is the sense of planning for re-use, of seeking to maximize return-on-investment on these data products? Why is it not yet routine that all data end up in managed repositories from where maybe they could be published to the web? in vendor-neural formats that guarantee their re-usability?
  • where are the instances data improvers being willing to return to the data originator the enhancements that they've made? Can we ever expect to do better than using the same old tired framework layers from DCW or whatever if no-one is prepared to feed back to authoritative custodians their improvements made to baseline products?
  • where are the mechanisms by which data improvements can be proposed or lodged? How many data distributors have procedures in place for dealing with feedback from their users when it comes to quality improvement?
Sorry to start the New Year with a rant like this but these issues are to important to not grasp an opportunity to highlight them. Fine we can talk melodramatically about "life and death situations in Kenya" and not be entirely exaggerating, but even the prosaic tasks of getting food distributed to camps on time, or ensuring that water is tanked from the most accessible clean sources, can help alleviate an otherwise completely miserable situation for thousands and thousands of people.

No comments: