A world that counts – the data revolution

Ecrire à Ban Ki Moon, c’est pour les adultes un peu comme écrire au Père Noël pour les enfants. On y croit plus ou moins, mais ça vaut toujours le coup d’essayer. En août 2014, Ban Ki Moon missionne un groupe d’experts indépendant pour l’aider à construire une vision pour les prochains plans dans le cadre des Objectifs du Millénaire pour le Développement, en focalisant sur les opportunités offertes par les données.

« The data revolution is giving the world powerful tools that can help user in a more sustainable future. The recommendations of the Group will be important inputs to the post-2015 debate and our efforts to shape an ambitious yet achievable vision. » (Ban Ki Moon).

Screen-Shot-2014-11-05-at-11.50.43-PM-820x1024Je retranscris ci-dessous le commentaire que j’ai laissé dans le cadre de la consultation du groupe d’experts ; et je vous invite à lire tous les autres commentaires ici : http://www.undatarevolution.org/measuring-sustainable-development/ , et le rapport issus des travaux du groupe : http://www.undatarevolution.org/report/

I see data revolution as the best lever to finally succeed in designing efficient policies, either at a national or at a local level. My leitmotiv is : you can’t manage properly what you don’t measure. (My comment is focused on development and help programs rather than on SDG)1. Data production, collection and thorough analysis is required to validate, at the field level, what programs work, and what programs don’t. To do so, those programs should be tested through statistically prooven experiments…. instead of just using examples for which the causes-consequences can easily be adapted to any conclusion one would like to raise. The work of Esther Duflo and her Poverty Lab at the MIT is exactly what should be done (in my opinion). It requires time and money, but it ensures efficiency and can avoid huge wastes.2. Data can help in raising awareness of the population on the as-is problems and difficulties they faced. Of course they know what difficulties they have to face, but they are not necessarily aware of the root causes of their difficulties, of the levers they could activate to improve their situation, of examples of what others have done to efficiently change their day-to-day. I’m far from thinking that those population should work on their situation completely on their own. Nevertheless, I’m sure that it can help a lot if they do want to be helped and if they do trust actions that are suggested to them. There is something that seems a bit paradoxal in the Development policies ; unlike problems such as racism, sexism, …. the one to organise the policies in the first place are not the one who suffer from the situation….3. The increase in the volume of data available, the increase of their quality and the increase of the analysis performance should help not only in identifying useful insights…. but also, and maybe mainly, on identifying new hypothesis that could be worth experimenting. Whatever the performance of the data analysis tools you use, if you ask the wrong question, you won’t be able to do anything interesting with the results (garbage in, garbage out). The exploitation of huge volume of data and the speed of analysis enable to test almost any combination of data to see if something emerges ; if so, you may find something to test through a robust experiment.

4. The increasing power of data analytics should help in taking into account and understanding individual situations instead of the average. The average is always true globally, but almost always false locally…. Development issues can be seen as a global problem, but I bet that efficient programs are programs that are specifically adapted to fit the situation of a local area and its specific context.

Easier said thand done…. and those are only my ideas, I do not have statistically proven experiments to support them 😉