By Bram van Rooij
Abstract:
The project ‘Sustainable Outlook’ uses the ‘RuimteScanner’ (Land Use Scanner) to calculate possible future land use. The goal of this information was to use it to support political and societal debates and to support policy-makers. The VisualScan project was carried out to explore the options of visualizing the 2D outcome of the LandUseScanner to a 3D visualization. Two approaches of visualizing this data have been developed. Though the outcome was never officially evaluated, in informal evaluations it became clear that the visualizations lacked a link with the cognitive map of the users.
The goal of this research is to communicate the Sustainable Outlook results effectively with visualizations in Google Earth. This research elaborates on previous research (VisualScan) and tries to improve the two ways of visualizing. This is done based on literature and – as said before – on experiences of previous research. Since the results have to be communicated effectively, a usability study is done in the form of a questionnaire.
There are a few basic factors important to create a usable and acceptable 3D visualization. A few who really pop out are ‘easy to learn’ and ‘good navigation and orientation’. Realistic visualizations are not preferable, since they can cause bias and misunderstanding. Also because the time-span of the visualizations (40 years) brings a lot of uncertainty. Changing the output of the Land Use Scanner from 29 to 11 classes has created possibilities for other kinds of visualizations, even though the MNP wants them in 3D. Google Earth gives a lot of advantages, because its standard features fulfill some of the user requirements.
Based on this information six different visualizations were made. A 2D visualization which showed only cells that changed and a 2D visualization that showed all the cells, regardless of change. In the same way two 3D visualizations were made. Also two abstract visualizations were made, one with symbols and one with pinpoints (standard Google Earth symbols). The usability of these visualizations is checked by a questionnaire.
The questionnaire was send to 26 persons, to which 10 persons responded. The 26 persons all came of the send list of Sustainable Outlook 2. The visualizations could be watched in Google Earth, the answers to the questions could be filled in on an online form. Six persons filled in the questionnaire, four persons e-mailed and did not fill in the questionnaire. This was mostly because of technical reasons.
The results of the questionnaire were as follows: The symbol visualization was preferred the most and was understood quite well. 3D visualizations were slightly preferred over 2D visualizations and visualizations that only show change are preferred a bit more than ‘normal’ visualizations. The pinpoint visualization was not understood well and was rated low. There seemed to be inconsistencies in the responses. Preferences of respondents differed in control questions from answers given to ‘normal’ questions. Responses seemed to be biased, an explanation for this has not been found.
The overall conclusion is that the visualizations are rated as quite usable. Whether the persons who were invited to fill in the questionnaire are representative is not clear, since the user-group could not be defined clearly. Usability is related to the user, so the usability of visualizations is depending on the user-group. Therefore the user-group should be clearly defined. The changes in output data (29 to 11 classes) will probably result in a design issue, since the classes are broader. It will be more difficult to visualize the land use types in a representative way. This could be solved be decreasing the level of realism of the visualizations.