humanitarian.info » Katrina

DisasterTech at eTech 2008

Paul Currion — Fri, 07 Mar 2008 18:09:54 +0000

The O’Reilly Emerging Technology conference is another one of these confabs that I watch from a distance, filled with a mixture of awe and dread. (Awe at the sheer brainpower that you can see in the many presentations, and dread at what might happen next.) This year, Jesse Robbins and Mikel Maron gave a presentation on DisasterTech to the poor and huddled masses that attended, updating some of their earlier thoughts on areas such as SMS, open source and distributed approaches.

Now I like Jesse and Mikel, and I agree with the lines along which they’re thinking, but when I see slides that say “225,000 deaths preventable with existing technology”, I start to worry about whether the expectations of those poor and huddled masses from the technology world are being raised just a little too high… but I’d rather wait until I can hear the audio rather than just read the slides.

One point they did make that’s worth picking up (because as far as I know, this is the first time it’s been explicit) is that disaster technology tends to follow this pattern:

Disaster
Ad Hoc Adaptation
Championship
Iterative Improvement

That’s pretty accurate – most of the more interesting developments of the last few years have followed that model. However there are definitely problems with Championship – not as a concept, but in terms of where to target your “championing” efforts.

I believe that most technology dissemination in this sector isn’t through organisational adoption – there’s just too much autonomy at the local level – but through word-of-mouth. Somebody sees something working, tries it out and takes it forward (or not) – but only within their particular part of their organisation. What this might mean is that we need to take a viral approach to this, rather than seeking to get management (i.e. top-down) approval, particularly if we want to reach smaller groups on the ground.

With that in mind, we have to be anthropologists as well as technologists. Cultures are different, communities are different and people are different; what “works” in terms of viral dissemination in the US is simply not going to work in Bangladesh. It’s not just the technology that has to adapt to these different environments; it’s us as well. Hopefully the start-up jockeys at the conference realise that…

Building an Emergency Operations Center on Groove and SharePoint

Paul Currion — Wed, 08 Nov 2006 12:38:32 +0000

Another ridiculously late notice on an article in TechNet magazine from October: Communication & Collaboration: Building an Emergency Operations Center on Groove and SharePoint. As always, mine eye is drawn to the lessons that can be drawn from these experiences, which the article sums up as:

First, network connectivity and bandwidth are not guaranteed… Second, the Internet isn’t always there. This may seem like a given in light of the previous constraint, but it’s important to highlight the potential frailty of an Internet uplink…Third, the definition of “users” expands greatly… How many organizations are ready to bring potentially hundreds of volunteers, contractors, and various civilian and military governmental staff into their information systems in a rapid and secure way? Finally, people need to access data using a variety of machines and connectivity scenarios. New users will often bring their own computers, configured in their own ways with their own sets of applications on them.

Now that’s like a primer for technology professionals on what to expect in a disaster. An EOC is a very specialised set-up – you’ll rarely find one during a disaster in a developing country – but the lessons run across the board. Wise words from John Morello.

Open Source Disaster Recovery

Paul Currion — Sun, 07 May 2006 21:36:59 +0000

First Monday publishes an article entitled Open Source Disaster Recovery: Case Studies of Networked Collaboration, a review of some of the initiatives that I referenced in my paper An Ill Wind? The Role of Accessible ICT following Hurricane Katrina. That paper has already discussed how these “non-traditional” voluntary efforts might fit into the broader picture of disaster response, but the First Monday article is the first time I’ve seen any discussion of how effective they might have been.

I spend enough time berating the UN and NGO community for focusing on outputs and failing to measure impact, so I’ll apply the same questions to the two projects covered by the article – Katrina PeopleFinder and the South East Asia Earthquake and Tsunami blog. The basic challenge for both projects was how to aggregate fragmented data streams; the key question for both was how to mobilise the enormous goodwill generated by the two disasters into useful resources.

With PeopleFinder, they certainly managed a huge amount of data processing and their site was clearly widely used (judging solely by the number of hits). Ethan Zuckerman characterised it fundamentally as a data entry problem in his blog entry on the history of the project, and that’s the sort of problem that’s amenable to the large scale voluntary effort that the PeopleFinder team put together.
However it’s hard to evaluate how many of those hits resulted in family reunification, and consequently it’s hard to judge whether all that effort was worth it. This is not exclusive to PeopleFinder – none of the sites running these services have been able to work out how useful they were – and to some extent it’s not relevant. The service needed to be provided, and they provided it – but it’s hard to make the case for replicating the service in other disasters if you can’t demonstrate utility.

The article cites an instance of a request for assistance by the Tamilnadu Tsunami Relief Initiative, and then goes on to excerpt a number of the responses. The article emphasises the diversity of the respondents – ministers from Scotland, graduates from Texas, psychologists from Denmark – and notes that the inititiative possibly suffered from differing expectations, conflicting agendas and high transaction costs. The article concludes that

Quantitative research investigating to what extent these volunteers actually used the blog to match their skills and resources with needs, and how good those matches were in practice, would help clarify the blogï¿½s effectiveness in identifying and allocating resources for disaster relief.

It’s a fascinating and useful article – but it’s just a starting point for these discussions, and David Geilufe has an excellent response on his blog. There’s still no real discussion of the impact of these initiatives, the actual relief they provided or facilitated to beneficiaries. Anybody following up on this line of research should not just focus on the emerging model of collaboration for disaster relief – which is clearly interesting in itself – but what it could actually deliver.

However one of the problems with this model is that everybody involved in them has day jobs, and once the immediate disaster is over, they go back to those day jobs. And blogging.

5 Years of Information Management for Humanitarian Operations

Paul Currion — Thu, 20 Apr 2006 16:23:35 +0000

Five years ago, I wrote an article for Humanitarian Exchange, published by the ODI Humanitarian Practice Network, entitled “Learning from Kosovo: the HCIC Year One.” It was the first thing I’d written about information management, and it was right at the start of my learning curve in this field.

I thought it would be interesting to update that article, and revisit some of the issues that it raised. So in the latest issue of Humanitarian Exchange, you’ll find another article, this time entitled “A Little Learning is a Dangerous Thing: five years of information management for humanitarian operations.”

I packed in more points than I probably should have, so it’s quite a dense piece, but it’s probably summed up by these two sentences:

Five years ago, I believed that better information management would enable better management overall. Yet it is hard to determine whether all this activity has actually improved the provision of humanitarian assistance, since there are no clear criteria for measuring their impact.

The good news includes more partnerships, sometimes between agencies, but also with the private sector; the bad news includes a lack of leadership in this area, particularly given the recent implementation of the cluster approach by the UN. The most important point I wanted to make, however, is one that I’ve mentioned before. I’m going to repeat it now, because it seems to me to be important enough:

We are public organisations, funded by public money, working for the public good. In the interests of accountability, all information that we gather in the course of our work should also be public. The only qualification that should be made to this is if that information might in any way endanger the safety or security of beneficiaries or staff. Without the free flow of information, the process of coordination is crippled, and we all must take responsibility for this.

So I guess that’s my new mantra.

Also interesting this month is an article by Jeremy Shoham of the Emergency Nutrition Network and NutritionWorks. The article, “Information is a prerequisite, not a luxury“ expands on the importance of information management in the context of Food Security. Although he makes the case for what I would say is an overly ambitious system, the article is an excellent overview of some current thoughts in food security (along with a jab at the bias towards food aid, which thankfully is starting to diminish).

Mapping Disaster Zones (Nature magazine)

Paul Currion — Tue, 21 Feb 2006 15:03:36 +0000

More map madness in Nature magazine, whose 16 February issue has a commentary piece on Mapping Disaster Zones, covering work done by the Global Connection Project in response to Hurricanes Katrina and Rita and the Pakistan earthquake. The entire article is worth reading, but I’d like to focus on some of the lessons they learned:

We learned many other lessons from Pakistan, not least the need to develop tools that can adapt to local conditions on the ground. Internet connections in Pakistan were often slow and patchy, making downloads difficult. Post-disaster feedback highlighted difficulties with printing out maps and locating specific settlements among tens of thousands.

That part just brings back too many bad memories, especially the time that all the ink cartridges in our map plotter cracked in transit. It made a very big mess and we had to clean it by hand. Everybody was very unhappy.

Another obstacle was the mismatch between the local situation and the way existing geographic data is organized and represented. For example, Pakistani villages do not have a single position, but exist along a right-of-way extending from the valley to a ridge top, migrating up and down based on the changing seasons. Thus assigning a fixed position to a village is wrong most of the time.

This is an interesting problem, and one that we come up against frequently in the field. (Jishnu Das has written some more about this, using Flickr to tag his photos from his field work, based on his work with RisePak and LUMS in the Pakistan response.) Not only are settlement positions changeable in some countries but they can move or disappear, whether due to earthquake or war. So what’s the solution? The article suggests that

All mapping efforts, whether community based or UN supported, would benefit from a stronger feedback mechanism, such as a community filtering system that enables authenticated individuals to update shared geographical data online.

That would be nice, but I don’t think it will ever work. Ground truthing is absolutely indispensable for any remote sensing, and that means feet on the ground. People actually working on relief don’t have time to log on to update data – even if they have access to the internet – they just wave the map in your face, shrieking “Your map is wrong!” Which means grabbing the map, making them mark the wrongness clearly, and updating the files accordingly.

As a result, a ‘community filtering system’ isn’t likely to work; you need an authoritative reference point, preferably in the field, a focal point for maintaining and disseminating spatial data. In the absence of government capacity to do this, a Humanitarian Information Centre usually does this, with UNJLC or UNMAC handling logistics and mines respectively; in the absence of any of those, we’re a bit stuffed, so any offers would be welcome.

The full article is at http://www.nature.com/nature/journal/v439/n7078/full/439787a.html.

[Update: More from Jishnu on Kathryn Cramer’s blog at http://www.kathryncramer.com/kathryn_cramer/2006/01/a_long_interest.html; Kathryn is also one of the authors of the Nature article, and her blog is worth checking out as well.

N-TEN Webinar – Hurricane Katrina – Innovative Information and Communications Response

Paul Currion — Thu, 29 Sep 2005 21:31:57 +0000

On Tuesday I participated in one of N-Ten’s online seminars, this one co-hosted by HumaniNet, Aspiration, and the Innovation Funders Network. The participants included some of the people who worked on the issues I wrote about in my Katrina paper, including Will Reed of Tech4All and some of the guys who worked on the KatrinaPeopleFinder data initiative.

It was fascinating to hear first-hand about the things I’d been reading about during the Hurricane. It was obvious that these guys were dynamic, but what struck me was that they were also very frustrated with the chaos following Katrina. Of course, that’s the nature of working in disaster response – if existing systems could cope, it wouldn’t be a disaster.

It was also clear that they had quickly learnt many of the lessons it took me years to pick up – particularly the basic truth that the best way to lessen the impact of a disaster and ensure quick recovery is to help communities to help themselves – and that’s where the technology comes in. We need to develop tools that help the “first responders”, sure – but perhaps we should also focus more on developing tools to help the communities that we’re supposed to be helping.

Beyond that initial interest, the webinar was slightly frustrating. It was a good way to get a lot of people around the table, but not so good when it came to talking about the issues. It was also difficult to establish where the common interests lay between the different initiatives – I was presenting Sahana, the Open Source development initiative that came out of the tsunami response in Sri Lanka – and it was hard during the seminar to really pinpoint how we can all work together.

N-Ten are an interesting outfit, and thanks to them for organising the webinar (although I’m still not convinced that webinar is a word that should be allowed to survive). I’m looking forward to meeting them when I’m in the States in October / November – they’ve invited me to be on the Panel Discussion at their upcoming conference, “Answering the call: Katrina, the Tsunami, Darfur, Afghanistan – lessons learned from the Global ICT responses” on October 17th. If anybody out there is coming along, I’d appreciate it if you didn’t throw popcorn when I’m trying to speak.

katrina on del.icio.us

katrina on del.icio.us

Paul Currion — Fri, 09 Sep 2005 23:16:18 +0000

Despite the terrible impact of Hurricane Katrina, it has been very positive to see the technology community using a wide range of tools to support the response. Particularly in areas such as family tracing and volunteer matching, many initiatives have been started privately, and resources such as Google Maps have really come into their own as a way of distributing information. I’m keeping a watch list of websites and articles on the use of technology in the Katrina response. You can find the (regularly updated) list on my del.icio.us roll, at del.icio.us/paulcurrion/katrina.

These initiatives are fantastic, and I’m loving the fact that so many of them are fairly spontaneous. The group that I’m involved with on the Sahana project are trying to field the prototype version, working with some volunteers in the US. I hope that we’re able to prove some concepts in this response.

However, two things occur to me:

1. Fantastic as these initiatives are, they are fragmented and unco-ordinated – much like any humanitarian response in the world, to be honest. However this is massively inefficient, and we don’t meet the needs of people and communities as quickly and appropriately as we should. Are there ways of rapidly developing network organisations to co-ordinate these initiatives, without destroying the volunteer spirit, spontaneity and inventiveness of the decentralised approach?

2. Despite the similarities in the response, this isn’t a humanitarian crisis like the ones I normally deal with. Aside from Kobe ten years ago, this is one of the few serious humanitarian emergencies that has taken place in an urbanised area in a highly-developed country. Many similar needs to an event like the Asia tsunami – housing, family tracing, victim identification, etc – but the operational environment is radically different – excellent infrastructure, functioning government, abundance of resources. At this point, it’s hard to see where to draw lessons – and easy for this event to skew funding availability and the development of useful tools for humanitarian responses in other parts of the world.

Baseline Magazine in May