Hunter and Goodchild explore current and future ways of visualizing error in GIS. An important distinction is made in the varied audience of these visualization tools, and how different representations should be used to target GIS novices and experts.
I am a big fan of structural guidelines built into databases to encourage or discourage certain practices. These prompts are fairly unobtrusive and stem errors at the data collection stage, pre-empting any wasted time on faulty analysis. The acknowledgement of variability of data from different data sources is increasingly important in an Internet filled with geospatial data.
The image below shows uncertainty in contour lines by the size of gaps in a contour; larger gaps represent greater uncertainty. To me, this visualization provokes deeper reflection on the limitations of the vector data model in representing a continuous variable such as elevation. However, this thought process is unlikely to be the case for GIS beginners. The eye is still capable of recognizing the general pattern and connecting the dots. Furthermore, it is confusing whether the gaps represent individual spots where it was impossible to obtain elevation data, or whether the entire contour line was measured with low resolution/precision.
Probability surfaces suggested by the authors are more concrete representations of the various possible outcomes. I believe it is important to show that analyses can still be derived, so that users are not discouraged by the increasing visibility of uncertainty and abandon the tools. A balanced message must be achieved to expose the limitations of GIS in a constructive way.
– Madskiier_JWong
Image Source: Pang, A. (2001) “Visualizing Uncertainty in Geo-spatial Data” http://www.spatial.maine.edu/~worboys/SIE565/papers/pang%20viz%20uncert.pdf