Most visualization techniques have been designed on the assumption that the data to be represented are free from uncertainty. Yet this is rarely the case. Recently the visualization community has risen to the challenge of incorporating an indication of uncertainty into visual representations, and in this article we review their work. We place the work in the context of a reference model for data visualization, that sees data pass through a pipeline of processes. This allows us to distinguish the visualization of uncertainty - which considers how we depict uncertainty specified with the data - and the uncertainty of visualization - which considers how much inaccuracy occurs as we process data through the pipeline. It has taken some time for uncertain visualization methods to be developed, and we explore why uncertainty visualization is hard - one explanation is that we typically need to find another display dimension and we may have used these up already! To organise the material we return to a typology developed by one of us in the early days of visualization, and make use of this to present a catalogue of visualization techniques describing the research that has been done to extend each method to handle uncertainty. Finally we note the responsibility on us all to incorporate any known uncertainty into a visualization, so that integrity of the discipline is maintained.
@InCollection{ brodlie:2012:RUDV, author = {Ken Brodlie and Rodolfo Allendes Osorio and Adriano Lopes}, title = {A Review of Uncertainty in Data Visualization}, booktitle = {Expanding the Frontiers of Visual Analytics and Visualization}, pages = {81--109}, publisher = {Springer Verlag London}, year = {2012}, editor = {John Dill, Rae Earnshaw, David Kasik, John Vince and Pak Chung Wong}, }