“Towards an early warning system: the effect of weather on mobile phone usage. A case study in Abidjan”

From Navigators

Revision as of 13:07, 7 May 2013 by Jcraveiro (Talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

João Pedro Craveiro, Fernando Ramos, Eiman Kanjo, Nour El Mawass

in D4D book. Mobile phone data for development, Vincent Blondel, Nicolas de Cordes, Adeline Decuyper, Pierre Deville, Jacques Raguenez, Zbigniew Smoreda, Eds.

NetMob, May 2013, pp. 18:1–18:11.

Abstract: In mobile networks, traffic activity within particular network cells follows predictive patterns. Relevant changes in these patterns may indicate a local problem (an emergency or any other sporadic event). Detecting these changes could therefore be used as an early warning system. The hypothesis we aim to address is whether we are able to detect unexpected changes in weather quickly, by monitoring changes in cell patterns. The motivation is the fact that it is usually easier and cheaper to prevent damage provoked by weather conditions than to reverse the damage. This is particularly relevant in the socioeconomic context of developing countries such as Ivory Coast. In this paper, and as a first step towards the development of an early warning system, we jointly analyse mobile data and historic records of weather conditions. We employ exploratory factor analysis to reveal latent variables in the weather data, and spectral analysis to exploit the periodicity of mobile and weather data (both independently and jointly). From these results we derive a model which, in spite of its current limitations, hints that our hypothesis may be viable if conducted with higher quality weather data.

Download paper

Download Towards an early warning system: the effect of weather on mobile phone usage. A case study in Abidjan

Export citation



Missing ResearchLine

Personal tools
Navigators toolbox