Virtual Networks and Poverty Analysis in Senegal
Do today's communication technologies hold potential to alleviate poverty? The mobile phone's accessibility and use allows us with an unprecedented volume of data on social interactions, mobility and more. Can this data help us better understand, characterize and alleviate poverty in one of the poorest nations in the world. Our study is an attempt in this direction. We discuss two concepts, which are both interconnected and immensely useful for securing the important link between mobile accessibility and poverty. First, we use the cellular-communications data to construct virtual connectivity maps for Senegal, which are then correlated with the poverty indicators to learn a model. Our model predicts poverty index at any spatial resolution. Thus, we generate Poverty Maps for Senegal at an unprecedented finer resolution. Such maps are essential for understanding what characterizes poverty in a certain region, and how it differentiates from other regions, for targeted responses for the demographic of the population that is most needy. An interesting fact, that is empirically proved by our methodology, is that a large portion of all communication, and economic activity in Senegal is concentrated in Dakar, leaving many other regions marginalized. Second, we study how user behavioral statistics, gathered from cellular-communications, correlate with the poverty indicators. Can this relationship be learnt as a model to generate poverty maps at a finer resolution? Surprisingly, this relationship can give us an alternate poverty map, that is solely based on the user behavior. Since poverty is a complex phenomenon, poverty maps showcasing multiple perspectives, such as ours, provide policymakers with better insights for effective responses for poverty eradication.