Communicating Uncertainty in Small-Area Spatial Change Estimates Using a Filter Approach Based on Probabilities

Peter Dorey, University of Southampton
Rhorom Priyatikanto, University of Southampton
Ho Man Theophilus Chan, University of Southampton
Natalia Tejedor-Garavito, WorldPop, Department of Geography and Environmental Sciences, University of Southampton
Chibuzor Christopher Nnanatu, University of Southampton
Amy Bonnie, University of Southampton
Eunice Mueni Williams, University of Cambridge
Andrew J. Tatem, University of Southampton
Carla Pezzulo, WorldPop, School of Geography and Environmental Science, University of Southampton

This paper introduces a methodology to visualise uncertainty in small area estimates of change in indicators derived from Bayesian model-based geostatistical (BMBG) methods. This approach is designed to support policymakers in evaluating subnational disparities and monitoring of indicators over time to assess the achievement of the Sustainable Development Goals (SDGs), aligning with the principle of "Leaving No One Behind" (LNOB) central to the 2030 Agenda for Sustainable Development. Using synthetic data, we show how posterior samples from BMBG methods can be leveraged to provide an interactive filtered mapping of the change in indicators based on the probability that the true value of the change in the indicator represents an increase or a decrease. This probability is derived from the distribution of the posterior samples. The visualisations can display mean values or more nuanced quantiles based on user-defined credibility levels. This filtering approach provides a robust yet flexible tool for displaying changes in indicators, facilitating effective communication of uncertainty to policymakers.

Keywords: Data visualisation , Bayesian methods , Data and Methods, Spatial Demography

See extended abstract.