At this point, VisWeek 2012 is just about halfway done, and there’s tons to write about, but I’ll try to keep these short by sticking to one day at a time. VisWeek now runs four parallel tracks for the best part of a week, so there’s no way I can tell you about everything that is happening out here in (today, surprisingly sunny!) Seattle. But I will tell you about what I think is cool. The usual caveats follow: omissions and mischaracterizations are all my fault.

VisWeek started on Sunday, and I saw Pat Hanrahan’s LDAV keynote on ``Divide and Recombine: an approach for analyzing large datasets’’. This was a nice tour through recent work (some from Hanrahan and his students, and much from other folks) happening on the big data community. Pat gave a nice overview of the reasons surrounding all the buzz around NoSQL databases (performance! flexibility! distributed computation!), and the challenges for visualization and data analyst researchers and practitioners: mostly, the toolchain is slow and clunky. And computer scientists are toolsmiths, so we should make better tools.

It was fun to realize that his talk was really about ``satellite issues’’ surrounding the topics which we traditionally consider visualization, such as: how can we get to the big data sources? How do we wrangle all that data into something usable for visualization, and how do we manage (computationally, and organizationally) all the analyses which we need to use to get meaningful visualizations? Much of his talk focused on R, the current darling language of data analysts. Interestingly, across the hall to this talk, a tutorial on ggplot2 and ggobi was being presented by Di Cook, Heike Hofmann (of whom I’ll have more to say soon). ggplot2 has been described as the hipster plotting library for R. But it really does kick ass. Of course, I’m partial to R since I work in the same building as some good folks who are deeply involved with it. Nevertheless, it’s nice to see the VisWeek community warming up to it, warts and all.

The other interesting observation Pat Hanrahan made was that problems which the (for lack of a better term) ``industrial big data community’’ tackle have a lot of overlap with the ones which we tackle at VisWeek. Somewhat inexplicably, there is minimal overlap of the communities: how many of us hang out at Strata, or the many great visualization and statistical computing meetups? It’s something to think about.

At the same time, someone mentioned to me tonight that in the user interface community, there is an analogous situation with UX vs. CHI. So is this simply an indication that visualization is now actually mainstream? In other words, is this fragmentation a symptom of success or an admission of defeat? I don’t know.

Plug: come see our panel!

I have a lot more to say about VisWeek 2012, but the bigger the wall of text, the more likely you are to go look at cute cats. So before you leave, let me plug a panel in which (because unfortunately Juliana Freire can’t make the trip) I will be presenting, about reproducible visualization research. This is a topic near and dear to my heart, so I am really honored (and more than a little scared) to be sharing the stage with Tim Dwyer, Tamara Munzner and Gordon Kindlmann. There’s few things I like better than a good argument, so this should be great fun.