A Lack of Communication and Visibility

Finding visualization projects and pretty pictures on the web isn’t exactly difficult, but what about actual research? What if you wanted to know what’s going on in visualization, and get a sense of what current work is the most interesting? There is no resource for this that I’m aware of, but there should be.

I’m currently in Washington, D.C., where I spoke at the DC Data Vis meetup last night (that was not my main reason for coming, but it worked out nicely). One of the questions I got afterwards was from a journalist who wanted to know how he could follow developments in the field. I didn’t have a good answer.

What We Lack

It’s a really interesting question because it highlights how bad we are as a field in communicating our research. Here are some ways you can currently follow what is going on in visualization. And yes, I will talk mainly about information visualization, but scientific visualization is doing much worse.

  • You can read the journals. TVCG, InfoVis, CGF, CG&A. Of course to do that, you’ll need subscriptions to all of them, or access to three different digital libraries (IEEE, Eurographics, and Sage). Not exactly ideal. And then, you just get all the papers, but no curation or context. Which of these are really new stuff, which are just minor improvements of existing work? Also, you’re missing all the relevant work in HCI, cognitive psychology, etc., but there is no way you’re going to read all those journals to pick out what’s relevant.
  • You can go to conferences, like IEEE VIS and EuroVis. That way, you’ll see presentations and you’ll be able to talk to researchers and people in the field. But aside from the time commitment and expense, each of these only happens once a year. What do you do the rest of the year?
  • You can read blogs. I’ve written summaries of conferences and the odd posting about particular papers (though mostly my own). But the conference summaries tend to only flag papers and give very little context. In addition, you can wait for Stephen Few to rip a paper apart, but he also only does that once or twice a year, thankfully.
  • Set a Google Scholar alert. Those are really hit-or-miss, though, and it’s very difficult to pick good keywords to match, so you don’t miss too much but also don’t get completely inundated with irrelevant stuff.

And that’s really it. At least, I can’t think of anything else.

A Visualization Research Blog

What do we need? Here are some thoughts for the ideal visualization digest blog/website/service.

  • Pick out interesting research. Somebody needs to make a judgment what work is worthy of discussion. That requires somebody with some knowledge of the field and its current directions.
  • Provide context. Just summarizing a paper is one thing, but pointing out similarities to and differences from other work make this much richer. This also requires somebody who knows the field.
  • Summarize and explain. The key is clearly to summarize the findings or ideas or techniques in a way that people can understand. That might mean writing, it might mean asking the authors for their prototype and using it on some data, etc. Making this worthwhile might require a considerable amount of work.
  • Keep feeding the pipeline. To keep a blog alive, you’d want to post at least one paper a week. You also need to have a pipeline set up, so that when the big VIS or EuroVis wave washes over you, you can fill it for the coming months, without overwhelming people with too many things at once.

There is clearly a lot of value in doing this, but I wonder who could do it. Most academics don’t care enough and don’t have the time. There probably needs to be a way for this to make money, and I don’t know if advertising alone would do that (probably not). Perhaps somebody can make this a paid service, but I have no idea if there is enough of a market for this (most media outlets aren’t exactly swimming in money).

I really hope that somebody will do this, but I’m not holding my breath. What I will do, starting next year, is discuss more papers – both current ones and older ones. I’m not committing to a schedule here, but once a month seems like a reasonable rhythm. We’ll see how this works out. There’s certainly a lot of interesting work out there that would deserve a broader audience.

Comments

  1. Brian says

    I think putting this together would be a real benefit, but also a real challenge. Practically speaking, perhaps the most likely to work approach would be an industry-academic partnership. Perhaps you could cast a somewhat wider net and include not just data visualization, but all visualization (cartography/mapping, scientific, etc.). See if you could find a set of sponsoring organizations, enough so the organization cost of sponsorship is relatively low, then try to get one or more universities involved as the source of labor. You would, I would think, want to target universities with good sized, and well-respected, visualization-related programs, especially graduate programs (as I figure grad students will probably be providing a lion’s share of the leg work ;-). They would need to have faculty members willing and able to participate in some of the comparative and selective work. Consider only peer-reviewed work, or have two categories, clearly separated, between peer-reviewed and non-peer work.

    Now, there are probably a whole host of potential ethical and political gotchas here, and the logistical issues are pretty daunting. That said, these kinds of groups are put together for many purposes, so it could be done. Whether you could get any industry buy-in absent some kind of economc benefit (e.g., competitive advantage to membership, disadvantage to lack of membership, reduced costs because a third party is now providing a service that used to be done by in-house personnel, etc.) is uncertain.

    Like the concept though.

  2. says

    What about starting a journal for that, including those recensions, comments, viewpoint articles and the like? It needs good peer review, but other reviewing standards than usual journals. I think other fields have an academic culture more open to these kinds of publications.

    I’ve written a couple of summary articles and conference reports for http://softvis.wordpress.com, but it’s really much work and there is no direct academic credit for that. Sad to say that, but writing another is usually not rated highest on my priority list.

  3. says

    It’s interesting that a journalist asked you this, because typically it would be *their* job to develop an information feed and resources for following a field they’re covering.

    But actually I am a big fan of experts doing more public-facing communication about their niches, and looking forward to your upcoming coverage – it’s also something I’m trying to do.

    I think you put your finger on the important question though: how to fund? I could see an academic outlet that has strength in both visualization and journalism collaborating to build an outlet using student labor (overseen by editors) to go the non-profit route. Or there are other models out there that seem to be successful for translating research into blog posts: http://crowdresearch.org/blog/

  4. says

    Good point, Robert. The visualization research should be more accessible people outside the scientific community.

    In addition to the points you mentioned, I can think of the following two points:

    1. Encouraging more surveys and STARs (State-of-the-Art Reports).
    They give a comprehensive overview of an interesting topic in visualization, the challenges associated with it, and guidelines on using available techniques or results for this topic.

    Some recent STARs on information visualization topics:

    Glyph-based Visualization: Foundations, Design Guidelines, Techniques and Applications (2013)
    http://diglib.eg.org/EG/DL/conf/EG2013/stars/PDF/039-063.pdf

    State of the Art of Parallel Coordinates (2013)
    http://diglib.eg.org/EG/DL/conf/EG2013/stars/PDF/095-116.pdf

    Visual Analysis of Large Graphs (2010)
    http://diglib.eg.org/EG/DL/conf/EG2010/stars/037-060.pdf.abstract.pdf

    Until now, Eurographics was the only conference offering a STAR track (and was more focused on Computer Graphics than on Visualization).
    Finally, EuroVis 2014 introduced a dedicated track for STARs
    http://eurovis.swansea.ac.uk/call-for-participation-state-of-the-art-reports.htm

    2. Encouraging the collaborative building of taxonomies

    I think taxonomies make it somehow easier for people to pick up papers and techniques about the problem they are trying to solve.

    here is an ambitious initative by Oxford for that: http://www.ovii.org/community/taxonomy/
    They have even a dedicated category for existing taxonomies http://www.ovii.org/community/taxonomy/132/
    But it is still in the beginning and needs engaging many people from the community to add their papers.

    There are also dedicated taxonomies on a specific topic:
    Hans-Jörg Schulz is running a website with nice and navigable collection of techniques for visualizing hierarchical data http://www.treevis.net/
    Christian Tominski and Wolfgang Aigner are running a similar collection of techniques for time-oriented data

    • says

      The problem with those STAR reports is that almost no one outside of a European institution can access them. Case in point, none of those EuroGraphics links work from my university connection.

Leave a Reply