This week we focused on information visualization with Niklas Elmqvist from the UMD iSchool. Niklas studies information visualization and human computer interaction. He joined UMD in the last year, after arriving from Purdue University.
For this week we also read Elzen & Wijk (2014), Heer, Bostock, & Ogievetsky (2010) and Heer & Shneiderman (2012). I enjoyed the two Heer articles because of their accessibility (they were written for the more general readership of ACM Queue), but also for their breadth. The 2012 paper in particular does a really nice job of summarizing a large number of visualization techniques by breaking them down into a taxonomy of data/view specification, view manipulation, and analysis process / provenance.
The surprise for me (since I’ve just been a dabbler in dataviz) is that the iterative feedback loop of the analysis/provenance piece is deemed an important part of the visualization itself. Niklas stressed this as he described how Visual Analytics which studies not only how to visualize data, but how the interaction between data processing, data visualization, computer interfaces and the human can enable new forms of reasoning that have previously been impossible, or at least very difficult.
The 2010 article was also very interesting to me because I recognized the name of Mike Bostock, who is a legend in the developer community for having played a part in the creation of Data Driven Documents (D3). D3 is a Web standards compliant data visualization toolkit. I have also used Bostock’s Protovis library, but learned from Niklas that Heer (his PhD advisor) also played a role in the creation of both Protovis and D3, as well as the Flare and Prefuse visualization libraries. It seems like there is a lesson here about persistence, or at least not staying still. Bostock was at the NYTimes until recently, helping bootstrap their data visualization capabilities.
We did spend a little bit of time talking about how essential it is to be able to share visualizations. We talked briefly about Bostock’s D3 publishing framework at bl.ocks.org, which allows GitHub repositories containing data and D3 visualizations to be easily published on the Web. I’ve heard from friends at the NYTimes that Bostock created a very similar in-house system for reporters and editors there.
I left this meeting more excited than I thought I was going to be about the propspects of learning more about data visualization. I hadn’t considered before how much of a HCI and data visualization problem there is lying in the web archiving domain. My immediate interest centers on the appraisal process itself: how do curators and archivists sift through social media to identify salient Web documents to preserve. But also the very act of exploring Web archives is quite under-developed. The Wayback experience of diving into the archive with a known URL and then wandering around in links is the de facto standard for Web archives. But how would search be presented: what are the new new and useful ways to search through time as well as text? It feels like there is a big piece of work that could be done in this area. At any rate, I definitely would like to take Niklas’ class when it is available next.
Elzen, S. van den, & Wijk, J. J. van. (2014). Multivariate network exploration and presentation: From detail to overview via selections and aggregations. Visualization and Computer Graphics, IEEE Transactions on, 20(12), 2310–2319. Retrieved from http://www.win.tue.nl/\textasciitilde selzen/paper/InfoVis2014.pdf
Heer, J., Bostock, M., & Ogievetsky, V. (2010). A tour through the visualization zoo. Commun. Acm, 53(6), 59–67. Retrieved from https://queue.acm.org/detail.cfm?id=1805128
Heer, J., & Shneiderman, B. (2012). Interactive dynamics for visual analysis. Queue, 10(2), 30. Retrieved from https://queue.acm.org/detail.cfm?id=2146416