For decades, uncertainty visualisation has attracted attention in disciplines such as cartography and geographic visualisation, scientific visualisation and information visualisation. Most of this research deals with the development of new approaches to depict uncertainty visually; only a small part is concerned with empirical evaluation of such techniques. This systematic review aims to summarize past user studies and describe their characteristics and findings, focusing on the field of geographic visualisation and cartography and thus on displays containing geospatial uncertainty. From a discussion of the main findings, we derive lessons learned and recommendations for future evaluation in the field of uncertainty visualisation. We highlight the importance of user tasks for successful solutions and recommend moving towards task-centered typologies to support systematic evaluation in the field of uncertainty visualisation.

No abstract

An essential purpose of choropleth maps is the visual perception of spatial patterns (such as the detection of hot spots or extreme values). This requires an effective and as intuitive as possible comparison of color values between different regions. Accordingly, a number of design requirements must be considered. Due to the lack of empirical evidence regarding some elementary design aspects, an online study with 260 participants was conducted. Three closely related effects were examined: the "dark-is-more bias" (i.e., the intuitive ranking of color lightness), the "area-size bias" (i.e., the neglect of small areas, since these are less dominant in perception than larger ones) and the "data-classification effect" (i.e., attention to data classification when interpreting spatial patterns). For each hypothesis, one or more maps in connection with single or multiple choice questions were presented. Users should detect extreme values, central tendencies or homogeneities of values as well as comment on their task solving certainty. In general, the hypotheses regarding the mentioned effects could be confirmed by statistical analysis. The results are used to derive conclusions and topics for future research. In particular, further comparative empirical studies are recommended to determine the best possible map types for given applications, also considering alternatives to choropleth maps.

Noise annotation lines are a promising technique to visualize thematic uncertainty in maps. However, their potential has not yet been evaluated in user studies. In two experiments we assessed the usability of this technique with respect to a different number of uncertainty levels as well as the influence of two design aspects of noise annotation lines: the grain and the width of the noise grid. We conducted a web-based study utilizing a qualitative comparison of two areas in 150 different maps. We recruited participants from Amazon Mechanical Turk with the majority being nonexperts with respect to the use of maps.Our findings suggest that for qualitative comparisons of 'constant uncertainty' (i.e., constant uncertainty per area) in thematic maps, noise annotation lines can be used for up to 6 uncertainty levels. During comparison of 4, 6, and 8 levels, the different designs of the technique yielded significantly different accuracies. We propose to use the 'large noise width, coarse grain' design that was most successful. For 'mixed uncertainty' (i.e., uncertainty is not constant per area) we observed a significant decrease in accuracy, but for 4 levels the technique can still be recommended.This article is a follow-up to our conference paper reporting on preliminary results of the first of the two experiments (Kinkeldey et al. 2013).

No abstract