During the last decades, digitization broadened access to cultural heritage collections for public audiences. Large online databases have been prepared for open access with simple search interfaces or visual exploration methods. In this position paper we discuss new challenges arising from these initiatives with regard to casual users. To meet their specific needs, we introduce a novel method for synoptic collection visualization which makes use of parallel space-time cubes to provide multiple spatiotemporal overviews, support free exploration, and to specifically engage casual audiences.
To facilitate sustainable access to our cultural achievements, cultural heritage collections provide
windows into the past and store data on their objects in multiple dimensions. Curiously enough, it is
the successful development of big cultural heritage databases like europeana.eu, which generates new
challenges of (in)accessibility: Existing user interfaces require prior knowledge about what is to be
found [
* Corresponding author
Cultural heritage databases aggregate massive amounts of digitized artefacts and associated metadata
and allow for queries according to the users’ prior knowledge on various metadata dimensions. Yet,
just as in museums, visitors often come without prior knowledge. Without a clear goal, they want to
interact with a digital collection in an exploratory way [
Visitors to digital cultural heritage collections need an orientation phase before they start
browsing or searching [
Museum research shows that only limited cognitive resources are available for learning
and exploration [
With exploration being the central activity in such settings [
We consider these specific needs of casual audiences to play a crucial role when it comes to the future acceptance and factual use of public interfaces to digital object collections.
While physical object collections are commonly explored in a close-up perspective (Figure 1, left) traditional means for overviews are floor plans, as well as lists, slideshows or grids in case of virtual collections. Several InfoVis methods have been proposed to widen the options for visual exhibition exploration. We distinguish between the visual encoding of spatial (i.e. cross-sectional, non-temporal) data aspects, and of temporal (i.e. longitudinal) data aspects (see figure 1, right), with “spatial” not only referring to geographic metadata, but also to their distributions in algebraic or vector spaces.
Visual Encoding of Spatial Data Dimensions: As place of origin counts among the most frequently documented data dimensions of cultural artifacts, geographic maps often serve as a standard visualization method to show the spatial distribution of artifacts’ origins. Given different thematic or stylistic classifications of cultural artifacts, set diagrams or treemaps offer insights into categorically or hierarchically structured data constellations. As for relational data (e.g. influences, references, inter-artifact relations) network diagrams or graphs enable users to explore the proximities and distances of artifacts or cultural actors in relational or topological spaces.
Visual Encoding of Temporal Data Dimensions: Especially in cultural heritage contexts, interfaces have to encode temporal information too. One prominent option to do so are linked timelines, usually implemented as a coordinated temporal view in addition to a spatial representation. Other prominent options are animation and superimposition with the former merging multiple temporal snapshots (often distinguished by different colors), and the latter mapping time to time. Another hybrid technique is the space-time cube (STC), mapping time to an additional spatial dimension.
As most of these methods have already been implemented in the cultural heritage data domain, we
focus on the specific needs of casual users: Which visualizations provide effective overviews while
keeping cognitive load low? Which methods provide conceptual orientation as an entry point and
support multiple ways of exploration, but also navigating between alternating views? With regard to
these questions, we will make the case for a more thorough consideration of space-time cube (STC)
representations (cf. figure 2), which have been shown in user studies to have unique strengths in
displaying multidimensional data [
First, the STC supports perceptual integration of multiple data dimensions: As a generic
spatiotemporal visualization method, the STC can display various spatial layouts on its data plane –
including maps, sets, and network graphs – and consistently map temporal information to its z-axis.
This genuine spatiotemporal layout principle provides “naturally” integrated views of three data
dimensions for synchronous perception, which conjoin into characteristically shaped point clouds,
clusters, flows, trees, or any other combination thereof (cf. figure 2). Experimental studies show that
the STC supports the identification of clusters or overall spatiotemporal patterns fast and efficiently
[
FFigure 2: Coordinated multiple space-time cubes, with a geo-temporal (left), a categorial-temporal (center), and a genealogical (right) layout, displaying the same data collection, subselection, and single artifact.
But the STC also allows higher dimensional integration: More than three data dimensions could
be perceptually integrated by either using other visual encodings (like color, size, shape, etc.) within a
STC, or by the InfoVis technique of coordinated multiple views, which also can be implemented as
coordinated, multiple space-time cubes (figure 2). The data planes of such parallel STCs can cover
different data dimensions and layouts while sharing the same selection of time. Further interactive
integration thus is available through the method of coordinated highlighting of selected data elements
or linking and brushing [
The STC can also enhance navigation: Bach et al. [
Based on these results we conclude that a visualization framework based on (multiple) STCs meets a substantive amount of design requirements for casual users. With regard to expected additional cognition and navigation support in connection with regular 2D views we consider its further evaluation in the cultural heritage data domain to be a productive research endeavor.
In this article we introduced an InfoVis approach to answer new challenges arising from the paradoxical (in)accessibility of cultural heritage databases for casual audiences. We found strong evidence that a multi-method interfaces based on the STC has the potential to meet essential design requirements delineated for casual audiences: They provide a synoptic overview on multidimensional collection data and enable users to generously explore collections from various spatiotemporal vantage points, while offloading data integration to 3D perception and keeping cognitive load low.
As a conceptual draft, this framework will be able to demonstrate its efficiency only in a series of user studies to come. Since implementations in the cultural heritage contexts have been missing until now, we expect a combination of comparative prototype evaluations to bring along new insights, about how different encoding methods (cf. figure 1) will perform with casual audiences. Possible implementation scenarios range from local implementations for visual collection exploration to extended knowledge communication initiatives in collaboration with various historically oriented disciplines, including Art History, Classical Philology, History of Science, History of Technology, Literary and Media Studies, etc. We further expect contributions to pedagogic and didactic methods development within these fields, i.e. on methods which support the multimodal teaching, presentation and collective exploration of historically oriented topics.
This research has been supported by the Austrian Science Fund (FWF), Project No. P28363 and the Austrian Research Promotion Agency (FFG), Project No. 835937.