Vincent Blondel's viewpoint
Vincent Blondel is a Professor of Applied Mathematics at KU Leuven in Belgium and President of the D4D Challenge. He conducts research on big data, specifically in the analysis of telephone data.
How can big data make a difference to agriculture in ACP countries and meet tomorrow's challenges? I
Big data can be mined to help bring about major! changes in agriculture and other fields. Overall it is felt that there is a dearth of information in ACP countries, whereas phone data is available -I the mobile penetration rate is very high in Africa! - as well as data from satellites and many other sources. Solar panels can, for example, be installed I on mobile phone masts as an inexpensive way! to determine the extent of solar radiation. This! gives us an almost minute-by-minute and quite! reliable picture of sunlight conditions and long-1 term patterns in different regions throughout a country. If you add hydrometric or rainfall sensors, one can calculate the efficiency of agricultural production. We are currently working with the World Food Programme on using mobile phone data and airtime credit purchases to estimate food security in Eastern Africa. The possibilities are immense.
What could hinder this development?
A key problem is that phone data is highly personal. Phone data can be very beneficial if it is exploited to generate, for instance, very concrete suggestions on how the Ebola virus spreads and thus how to enhance its control, whereas extremely sensitive data may also be disclosed, even if they have been rendered anonymous, and I aggregated on a large scale. A reasonable balance must be found between the protection of privacy and the accumulation of sufficiently rich data that could be processed to benefit society.
Are there sufficient human resources in Africa to tap these data?
Public research centres and universities exist. However, it goes without saying that the most advanced scientific expertise in big data is in North America and Europe. Relevant information, however, can be processed and extracted from these data using basic tools, without having to rely on the latest advanced techniques. Today, the problem is in boosting awareness and promoting ways to exploit the data from big data.
