Hacking Open Data on Development Aid

FrontlineSMS founder Ken Banks wrote an interesting blog recently asking why so many of us technology for development (ICT4D) types are content to work remotely.


Ken’s point was that just because technology makes it possible for us to work remotely it doesn’t make this the most efficacious way to work. A commercial company wouldn’t normally make an investment in product development without first doing significant market and customer research, yet in the ICT4D field it is not at all uncommon for techies to develop cunning technology solutions and then go out in a vain search for a problem to solve with it. Involving users in project design and development is not just good business sense, it’s an essential part of doing international development well.


In August, I went to the Development Data Hackathon at The Guardian - great people, great ideas, and unlike most of the hack events I’d previously attended it was not numerically dominated by spotty white men; there was real diversity among the participants, who brought a wide range of relevant experience and skills to the table.


For the uninitiated, a hackathon is an event where a bunch of people volunteer to turn up at a particular time/location to see if they can hack together a solution to a problem. The hackathon at The Guardianwas devised to see if we could use public information recently released as open data to create new knowledge/solutions for international development challenges.


One enterprising team hacked together a map locating all of the known clean water sources in a particular province of South Sudan using data released by the United Nations Development Programme (UNDP). A second team mapped geo-tagged data on aid by sector in Malawi (hover over the circles to see who’s making their data open). Yet another team worked on trying to understand the relationship between media coverage of humanitarian emergencies and aid donations using information scraped from news sites such as BBC, CNN and Channel 4.


Ingenious as these hacks are, they also reveal how we view aid from the perspective of development actors in London rather than from the perspective of people in rural Africa – we set ourselves managerialist problems. It is an understandable consequence of running a hackathon in London that technical fixes are developed without any real input into the setting of priorities or into the research and development process from the intended “beneficiaries” of development. This is true to some extent even if the hackathon takes place in one of the 50+ technology and innovation hubs now active across the African continent. Barely a week seems to pass without a new mobile apps competition or hackathon taking place but perhaps there is a need to hack the format of these events?


“We wanted to trace the money allocated for rural development to discover what percentage is actually spent in rural communities in Africa, but no data has been released in a format and sufficient details that would make such transparency possible”.


If development is conceived of as a geek event that takes place in the capital city how can it practically be informed by the perspectives and priorities of those marginalised communities that it is intended to benefit? Paulo Freire, who worked for many years on agricultural reform in Chile, said that it was important that the intended beneficiaries of development should be lead participants in the conception and production of innovations. He argues that the intended beneficiaries of development should at the same time be its authors and feel ownership of the process. This serves equally well as a definition of agile software development and of human development.


I have been working in the field of ICT for development (ICT4D) since 1988 and technology has a positive role to play, but I still have to struggle against the urge to put the technology before the people processes. I know that hackathons can and do result in ideas that can make a positive contribution – such as the M-Farm online marketplace, for example – but there is room to improve the format so the voices, innovative ideas and development priorities of the intended beneficiaries of development, are better reflected and better served by such events. I don’t have the answer, but I am confident that if we put the right heads together we could hack a solution.


Incidentally, the team that I was in at the Open Data Hackathon was trying to map the flow of millions of dollars of public funds from the World Bank and other agencies. We wanted to trace the money allocated for rural development along its route, via various international funds, development agencies, government departments and implementing organisations to discover exactly what percentage of the initial amount is actually spent in rural communities in Africa. And – you will never guess what – no data has been released in a format that would make such transparency possible.



Where the data represents the accounting of public funds derived from tax revenues, as is the case with the World Bank, United Nations and the UK government’s Department for International Development funds, or where the funds are derived from public appeals, there is really no defence for not publishing how our money is spent in a format that makes this analysis possible.


Only when agencies release data in a format that makes it possible to produce this level of accountability will it be fair to say there is a genuine commitment to openness and transparency in aid spending.


N.B. This post originally appeared in my column in Computer Weekly.