Abstract:

Data visualization provides users with intuitive means to interactively explore and analyze massive datasets, which can be dynamic, noisy and heterogeneous, enabling them to effectively identify interesting patterns, infer correlations and causalities, and support sense-making activities, making it possible to amplify human cognition.

Cultural Heritage (CH) institutions are a great source of high-quality data, that is not really being tapped into by data scientists. It is important to understand, however, that cultural data are different from those provided by natural sciences. They are seldom discreet and univocal. Regarding location and time, they are also heterogeneous, given with different granularities, and, for some cases, present uncertainty.

The vision of ClioViz is to take advantage of current technologies to extract the best possible information from huge datasets and synthetize them in visual forms that combine scientific visualizations and interactive maps. Data analysis will be used to retrieve meaningful information from datasets in the field of CH, with special focus on the visualization of the multispace and multitime complex variables, benefiting from visualization strategies coming from GIS and scientific visualizations.

Within the scope of the ClioViz project, we have developed a gallery of graphs (e.g., barplots, violinplots, spiderplots) that aims to offer visualization of the uncertainty of location for some heritage objects. This paper explores the understanding of such visualizations from a case study with undergraduate students: on the one hand, students in the field of data science; on the other hand, students in the field of art history. To that end, we have conducted a pair of workshops where the aims and scope of the ClioViz project were explained to the students, and then they individually fulfilled an online questionnaire to evaluate the understanding and/or usefulness of the proposed visualizations. Results showed that most of the students understood more straightforward visualizations such as heatmaps, scatterplots or barplots and considered them as optimal to explain the uncertainty of locations. However, more complex visualizations (violins, spiders, etc.) were considered confusing, by mainly the art history students. Students also proposed ways to improve the understanding of visualizations, such as changing some of the aesthetics and adding interaction to the graphs. Based on their feedback, the graphs have been improved.

Keywords:

Data visualization, workshop, multidisciplinary, cultural heritage.