2020/05/25 by Taylor, Aimee R, Watson, James A, Buckee, Caroline O
#Applications (stat.AP) #FOS: Computer and information sciences
paper · doi:10.48550/arxiv.2005.11993
Maps have long been been used to visualise estimates of spatial variables, in particular disease burden and risk. Predictions made using a geostatistical model have uncertainty that typically varies spatially. However, this uncertainty is difficult to map with the estimate itself and is often not included as a result, thereby generating a potentially misleading sense of certainty about disease burden or other important variables. To remedy this, we propose simultaneously visualising predictions and their associated uncertainty within a single map by varying pixel size. We illustrate our approach using examples of malaria incidence, but the method could be applied to predictions of any spatial continua with associated uncertainty.