To content
Department of Spatial Planning

Modelling Twitter response to rainfall using urban mobility flows

Maps showing the Twitter response to rainfall in relation to population and income © de Andrade et al. (2021)​/​IJGIS

Although it is known that urban inequalities can lead to biases in the production of social media data, there is a lack of studies that assess the impact of intra-urban movements in realistic urban analytics scenarios based on social media. Our recently published study examines the spatial heterogeneity of social media data with respect to regular intra-urban movements of residents using a case study of Twitter activity related to rainfall in São Paulo, Brazil. We apply a spatially autoregressive model that uses population and income as covariates and intra-urban mobility flows as spatial weights to link different areas. We use this model to explain the spatial distribution of social response to rainfall events in Twitter as compared to actual rainfall radar data. The results show a large spatial heterogeneity in social media response to rainfall events associated with intra-urban inequalities. Our model performance (R^2 = 0.80) shows that urban mobility flows and socio-economic indicators are significant factors in explaining the spatial heterogeneity of thematic spatio-temporal patterns obtained from social media. Therefore, urban analysis research and practice should consider not only the influence of the socio-economic profiles of neighbourhoods, but also the spatial interaction caused by intra-urban mobility flows beyond the night-time census to explain the spatial heterogeneity in the generation of social media data.

Full reference to our work published in the International Journal of Geographical Information Science:

de Andrade, S.C., de Albuquerque, J.P., Restrepo-Estrada, C., Westerholt, R., Morales Rodriguez, C., Mendiondo, E. M. and Botazzo Delbem, A. C. (2021): The effect of intra-urban mobility flows on the spatial heterogeneity of social media activity: investigating the response to rainfall events. International Journal of Geographical Information Science, volume and issue pending. DOI: 10.1080/13658816.2021.1957898.