Some people thought that my previous post on needs assessments implied that I’m really unconvinced by open data. Surprisingly that’s not the case – I know, even I was surprised that my curmudgeonly instincts failed to kick in – I’m pro-open data, although I’m typically European in being sceptical about the scale of the potential benefits. The reason that I’m pro-open data is that it’s increasingly clear that governance systems in general are unable to cope with the complexity of post-industrial organisation. We need new forms of governance, we can’t rely on our governments to generate those forms themselves, and the decisions of an informed public are the only possible source of legitimacy.
Having said that, open data in government 2.0 [Wow, has that term dated quickly.] does not necessarily mean the same as open data in the humanitarian sector; importing of principles and approaches wholesale from one sector to the other is almost never effective. In more practical terms, if governments are already signing up to release their data in general, that already covers data that might be deemed humanitarian specifically. There is almost no data which can be labelled solely ‘humanitarian’ [most of it relates to population movements] and there’s an increasing amount of ‘development’ data online. Since we’re already seeing a big push in the open data movement globally, I’m content to let that play out rather than try to shoehorn it into the humanitarian sector.
While we are seeing is movement in open data on the development side, that seems to have translated into more transparency regarding development finance, rather than a push to turn everybody into Hans Rosling. Rosling’s work is fantastic, and has done a lot to raise the visibility of development issues, it’s not intended to provide serious tools for decision-making at the national or regional level; the data simply isn’t detailed enough. That’s changing, and will continue to change, as the Age of Big Data bears down on us; but Big Data doesn’t resolve policy issues, and technocrats are often misguided in their expectations.
So what do I think open data can achieve? Simply put: more effective assistance to disaster-affected communities. What sort of data do we need to achieve that? Equally simply: operationally relevant data. I don’t care about data being shared as much as I care about data being collected; you can’t open up data that you don’t have, so our first priority should be getting decent data in the first place. The problem at the supply end is the lack of systematic data collection, which needs to be addressed with better survey work; and data should be collected solely in order to facilitate better decision-making – there’s no other reason to collect it. If that data isn’t being used for decision-making, then sharing it is irrelevant, which means that advocacy for and training in using data for decision-making is a higher priority than making data open right now.
The next consideration is that there’s always a cost to sharing data: although that cost has become substantially lower (great job, internet!), it still exists. Do the benefits of sharing data outweigh the cost? Please don’t argue that the costs are neglible, because they’re not – it’s a massive investment in an organisational change process just to get on the first rung of data sharing, there are lots of technical costs, and the potential for disaster is high (as the UK National Health System can tell you). I think the benefits outweigh the costs, but we need to build a stronger case than currently exists, which means we need to do a better job of drawing the dots between data and impact.
And this is where it gets tricky. Government 2.0 is all about transparency and accountability, and Development 2.0 is an extension of that, with an additional component of aid effectiveness. I would argue that this effectiveness component is much, much more important in Disaster 2.0 (ha!), but there’s also a tension around the role of government that a lot of people are in danger of missing. Hopefully everybody can agree that the big push for better data needs to be at country level – but humanitarian organisations frequently have a more adversarial relationship to government than UN agencies. We can’t expect MSF to participate in open data sharing with the Government of Sudan.
This isn’t just a question of context, this is a question of ensuring that humanitarian principles are preserved in any data sharing agreement: for example, if neutrality is in tension with transparency, which principle wins out? That’s a question that the current discussion around open data has simply not addressed, but it may turn out to be the most important question of all.
Related posts:
Following a brief discussion of this post on twitter, and a request, moving some discussion here.
Although I work more in what we’d call “development” than humanitarian assistance I liked this post, and it has some important insights. I particularly like the point that a lot of the data we come across online which tells good stories is not actually that useful at an operational level. And I agree that a lot more work is needed to improve “operationally relevant data”. In my field, HIV prevention and treatment, this would be data that tells us who is providing what types of service, to whom, and where, and for how long, so that we can figure out what gaps there are. There is a particular challenge though, which is that while a lot of clinical services are fairly easy to define, a lot of HIV prevention and care work happens outside of clinics, generally led by NGOs, each with slightly different approaches. Quite a lot of work needs to be done to better classify and describe these interventions. I suspect the problem isn’t specific to HIV either.
One issue raised on Twitter was that the process of getting data/evidence into decision making is not linear. But I think that it is important to differentiate data and evidence (which I think you do in your post by the way). I’d agree that getting evidence of “what works” into policy is generally very complex; but I think the point in your post is more about making good data on operations available.
Thanks for the positive feedback, Matt. You’re right, the problem of definition around Who’s doing What Where is not unique to HIV – it’s a long-standing problem which hasn’t been successfully addressed. I think one reason for that is that there are conflicting requirements for that data, which means that it’s complicated to collect it and (equally) complicated to present it in a meaningful way.
Pingback: You want capacity? We got capacity. Possibly. | humanitarian.info
I would like to support the call for data transparency in humantarian assinstance. Only real-time field relevant data will be able to steer the different actors in chosing interventions that are not overlapping and complementary. Indeed: much of what we now call “coordination issues “, would be solved with better data sharing on who does what, and an accountability on results.
Even if confidentiality of data is maintained in all cases where it is deemed necessary, the impact of data gathering and sharing would be huge. So it would not be a good policy to do nothing until we have solved every issue.
Meanwhile, I too am afraid fo the cost of the whole venture. However, this must be compared to the cost of coordination as it is practiced nowadays. A cost that is never made explicit, but possibly too high already.
Sam – I agree with nearly everything you say, including the idea that a lot of “coordination issues” would start to be solved with better data sharing. One thing that does worry me is the idea that transparency will inevitably lead to improved performance, rather than just to raised expectations.
Pingback: A brief exploration of Open and Big Data: From investigative journalism to humanitarian aid and peacebuilding | ICT for Peacebuilding (ICT4Peace)