The UN high-level panel recently called for a ‘data revolution’ to underpin the planning and evaluation of the post-2015 development goals.
A taste of what might be in store was presented at the European Development Days meeting in Brussels, Belgium, this week (26 November).
Mobile communications company Orange has huge amounts of data on traffic between its mobile masts, and in 2012 it decided to release some of that data from Côte d’Ivoire to researchers worldwide in a challenge to make use of that data. It anticipated 40 or 50 project applications and got 260 instead, though few came from African countries or Côte d’Ivoire itself.
The results, first released in May this year, were astounding. What had been — and largely remains — locked-up data, was analysed and presented by scientists to reveal a host of information with development applications.
Examples include better understanding of disease epidemics and their control, better traffic and parcel delivery planning, and insights into social divisions.
Releasing that data to scientists costs next to nothing, according to Orange’s vice-president of marketing vision, Nicolas de Cordes. But the collective effort researchers put into analysing it made it useable and thus added up to ten million Euros in value, according to rough estimates presented at the meeting.
And this is just the tip of the iceberg. Orange is now in talks with universities and ministries in Côte d’Ivoire about following up with a closer analysis of some of that data. It is also planning to repeat the experiment in another African country though it won’t reveal which one yet. And there is so much other data it has stored away, from, say, the age of mobile-phone users to information on solar radiation intensity at its solar-powered masts across the country.
A sustainable revolution?
No one yet knows if and when other telecoms will follow Orange’s example or whether the company will keep making its own data available for free. The experiment shows it could be a win-win situation with clear development benefits, the meeting heard, but it also carries potential loss of revenue for Orange — as it could always sell the data instead.
It is still too early to say if such a sway away from open data would negatively affect their use in development. One thing seems certain: any use is better than keeping the data shelved on servers.
And while the picture painted at the European Development Days was mostly rosy — of a revolution springing into action — caveats were heard too.
For example, the data for development movement is driven from the West, as are other zeitgeisty initiatives such as Corporate Social Responsibility, as one delegate noted in a session on the topic’s deployment in Africa. But do Africans actually want or need such schemes?
The research teams that applied for Orange’s Data for Development Challenge were mostly from the West, with only a few coming from African countries such as Cameroon and Senegal, even after the deadline had been extended and the challenge re-advertised specifically in Africa.
A possible solution would be to make future calls for research projects conditional on collaboration with an African team, to build capacity there too, remarked a delegate form the British Council, which uses such a strategy in its own projects.
Another challenge is involving official statistical offices in Africa, which should be supported to build their capacity through such initiatives, the meeting heard from a representative of PARIS 21, an initiative that works to see statistics and data better used in the developing world.
But Orange tried to do this and faced a challenge in doing so. Its approach might have been seen as a threat by the Ivorian stats office, which gets funding to do research with traditional methods, the meeting heard. Big digital data can sometimes provide the same information for much less money, and the challenge is to get the stats offices to see such initiatives as an opportunity rather than a threat.
Finally the data needs to be understood in context. Orange is only one out of several telecom providers in Côte d’Ivoire and its user data only account for roughly a third or so of all mobile communications there. In addition, many poor people and children do not own a mobile phone. All this means that any conclusions drawn from such data analysis at the moment come with a big warning sign, and researchers need to take all these caveats into account when making sweeping conclusions or even policy recommendations based on incomplete knowledge.
> Link to Orange's D4D Challenge
> Link to all submissions to D4D Challenge
See below for a podcast from the session: