User testing of dynamic geovisualizations: lessons learned and possible improvements for cartographic experiments

ABSTRACTThis article deals with the results of a user study, which aimed to compare two geovisualization interfaces (one interactive and one animated and interactive) displaying the chain of events during past major floods. Some limitations were identified concerning the variables measured (accuracy and speed of answer) and the tasks to be performed by users. Results have shown that the two common metrics used in map evaluations seem not to be sufficient to reveal differences between map types. Other metrics extracted from the participants’ interaction logs have been used successfully. Finally, to decrease the variability between the results of experiments and to make them more comparable with previous tests, one solution might be to develop and share reference data sets and sets of tasks in several themes, with known cognitive complexities, as is already done in other research fields.