New Approaches to In-Browser Data Visualisation

Canberra.
The final speaker at DHA 2012 this afternoon is Mitchell Whitelaw, whose interest is in data visualisation. The work he’s presenting here builds on the Prints and Printmaking Australia Asia Pacific database, and Mitchell’s project has explored the opportunities to present these data (20,000 works, 4,000 artists) in new, visual ways.

Mitchell has done similar work for other national archives collections, as well as for Flickr Commons image datasets. Such work provides exploratory visualisations of cultural collections, so far mainly with applications built in Processing and Java – but those technologies don’t work very well in browser environments, so recently this work has shifted more towards in-browser visualisations (using HTML and Javascript).

Mitchell is now presenting an interactive user interface which does this for the artists covered by the Prints and Printmaking database, and enables users to browse these artists and their works. A second interface takes the works as a starting-point, and presents the works arranged by decades; it also breaks down those works into different print types. (This also enables users to quickly form and test hypotheses about the collection.) Finally, a third interface provides a more local point of access: it begins with a single artist and enables users to browse all of the artist’s work, as well as their relationships with other artists and collaborators; users are then able to skip from artist to artist.

This, then, combines both visualisation and Web design; there no longer is an inherent technological distinction between these two fields, but they are collapsing into one another in modern browser environments. The interfaces presented here are generous in the information they provide (and counter-distinguished from ‘stingy’ search interfaces which provide only the key matching records); they show everything, or at least a lot, and manage to do so by providing clues or samples (carefully selected partial information) which add to the richness of the exploration. The key approach here is to show different scales of information (micro, meso, macro) at the same time, to provide deep insight as well as broader contexts.

It’s also important here to provide a number of separate, linked tools for exploring the dataset, enabling the user to navigate between these different views rather than forcing them to accept one of those possible views as the only option. This still is a technical challenge for current browser technology (as well as for the APIs it builds on, which are often still geared towards search), and therefore represents an area which is still rapidly developing.