Abstract

This seminar explores the challenges of analysing SMS data from interactive radio shows in local languages. We focus on a pilot aimed at understanding the collective beliefs of radio audiences in Central Uganda related to complications during and just after pregnancy. Combining contextual knowledge with data science and social science, we identified the perceived causes of pre-eclampsia and gender differences related to the locus of causality. The potential of applying our tools and methods to analyse textual data in low or no resource languages will be discussed.

Africa’s Voices Foundation provides detailed insights into the views and priorities of remote, under-served and bottom-of-the-pyramid social groups mainly in East Africa. We work on behalf of social service providers and policymakers to engage communities in inclusive ways on broadcast and new media, such as through local language radio discussions. We then analyse the multitude of conversational data for insights into the community’s views and opinions. This unique service enables organisations to build a deeper understanding of target groups, which can be used to design services and evaluate effectiveness.