WGBH, Boston’s PBS station, producted over a third of the nationally broadcast PBS television programming, and therefore archives hundreds of thousands of hours of video content, still images, and related documentation. It’s a huge and valuable archive, but finding anything inside of it is a nightmare. In 2006, they launched the “Open Vault”, a searchable online site highlighting the contents of the archive, but quickly found that scholars need a completely different set of tools. With help from teh Andrew W. Mellon foundation, they developed a new small-scale prototype and put it in front of users, with great results.
The lesson here is that, for our scholarly audience, the attractive yet random mosaic was “pretty” but not substantive enough. Yet the working relationship map, while it has the possibility of being very interesting, is not intuitive as currently implemented. We need to spend more time striking a balance between utility and usability for this feature.
All of these visualizations are made possible by the depth of cataloging we have for our prototype record set. We have cataloged to the item level and sometimes have gone farther to tag people, places, dates and topics within the content of a record. In the Robert McFarlane example above, a PBCore record describes the interview media asset, while a TEI-encoded transcript provides the links into the video content itself.
Other visualization possibilities we can explore with this depth of data include tag clouds, maps and timelines. We are looking into placing records in context along a timeline of life or collection dates that would look something like what is illustrated by Figure 5.
The entire paper, with examples of the visualizations they produced and results, is available on their site.
