Abstract: This paper focuses on how computer graphics and visualization can help users access and understand the increasing volume of geo-spatial data. In particular, this paper highlights some of the visualization challenges in visualizing uncertainty associated with geo-spatial data. Uncertainty comes in a variety of forms and representations, and require different techniques for presentation together with the underlying data. In general, treating the uncertainty values as additional variables of a multivariate data set is not always the best approach. We present some possible approaches and further challenges using two illustrative application domains.

As these components become more affordable and widespread (e.g. GeoSensor/Dust Networks), and as the volume and richness of geo-spatial data set being collected increase, the need for visualizing these data in an informative and consistent manner become more acute. In particular, there will inevitably be more concern about the accuracy, timeliness, and confidence of information being displayed – specially if the data are coming from multiple sources, or by their nature of collection contain some inherent uncertainty.

Depending on the domain, uncertainty can be associated with variability, but also sparsity in data, uncertainty in the model. It can arise from measurements, registration, and calibration operations, but also from processing of the data themselves. Uncertainty my be represented in different ways such as scalar, intervals, tuples, or distributions at each geo-spatial coordinate. Different visualization techniques must be developed to present these uncertainty representations together with the underlying data.

There is not consensus or universally recognized meaning for uncertainty in the literature. Uncertainty is a multi-face characterization about data, whether from measurements and observations of some phenomenon, and predictions made from them. It may include several concepts including error, accuracy, precision, validity, quality, variability, noise, completeness, confidence, and reliability.

It try to model (in process) the different components of spatial uncertainty

A definition of spatial uncertainty

In [1], spatial uncertainty is defined for both attribute values and position. It includes accuracy, statistical precision and bias in initial values, as well as in estimated predictive coefficients in statistically calibrated equations used in the analysis. “Most importantly, spatial uncertainty includes the estimation of errors in the final output that result from the propagation of external (initial values) uncertainty and internal (model) uncertainty.”

Visualization challenge
Uncertainty is an important issue with geospatial data sets, and hence it is not surprising to see a large number of papers from related fields such as geography and cartography. There is more than one way to classify how uncertainty can be visualized. One is by how uncertainty itself is represented, another is by how uncertainty is encoded into visualization. For the latter, there are two general ways of combining uncertainty into a visualization:

• mapping uncertainty information as an additional piece of data. Incorporating uncertainty information into the visualization by mapping it as transparency, haze, blur, etc. to alter the appearance of the underlying data.
• creating new visualization primitives and abstractions that incorporate uncertainty information. Modifying the visualization primitive itself so that uncertainty can be encoded with the data, and usually in such a way that the interpretation of both data and uncertainty cannot be visually separated.

Relation to my thesis: The source of data uncertainty are source errors, process errors and use errors [2]. In many cases uncertainty in geo-spatial data is as important as the data itself. My research focuses on how best to communicate uncertainty information accurately and effectively. This papers offers a state of the art on visualization techniques that treat uncertainty as an integral element with the data. I am deeply interested in how spatial uncertainty is handled in traditional domains such as geography and cartography. Other fields with works similar to my interests are information visualization [1] [2] (displaying uncertainty), artificial intelligence (reducing uncertainty), robotics (wayfinding with uncertainty), psychology (coping/dealing with uncertainty, spatial cognition), and complexity theory (contextualizing uncertainty).

References to read:
[1] H. T. Mowrer and R.G. Congalton, editors. Quantifying Spatial Uncertainty in Natural Resources: Theory and Applications for GIS and Remote Sensing. Ann Arbor Press, 2000.

[2] M. Kate Beard, Barbara P. Buttenfield, and Sarah B. Clapham. NCGIA research initiative 7: Visualization of spatial data quality. Technical Paper 91-26, National Center for Geographic Information and Analysis, October 1991. Available through ftp: ncgia.ucsb.edu. 59pp.