The Instruction and Outreach Department manages and coordinates library research instruction for students, faculty and staff through course-related workshops, outreach activities, personal consultations, research guides and other instructional materials.


Our blog has moved!

You should be automatically redirected in 6 seconds.
If not, visit
http://blogs.library.duke.edu/dukelibrariesinstruction/
and update your bookmarks.

Sunday, November 14, 2010

Data Visualization is Pretty

Even before library school, as a paraprofessional working in a library, I felt like I’d been hearing about “data visualization” and libraries for years. When I hear about it, sometimes it’s in that irksome, next-to-meaningless way (“Data visualization is something that is increasingly important to the librarian’s toolkit”); sometimes it’s in a provocative way (“Visualization is the only way to show the true beauty of statistics”); sometimes it’s in an offhand/amusing way (“Data visualization tools make things pretty”); and, yes, sometimes it’s in a guilty way (“Oh, yeah, um, data visualization. I really gotta learn about that... sometime”). Because I’ve been hearing and reading about it for so long, I feel like I’m way behind my colleagues in exploring the wide world of data visualization, even if that still means being ever-so-slightly ahead of some university library users. Recently, I decided to just jump in and see what I could handle. And it turns out, there are actually a handful of fun, potentially useful things that don’t require much specialized technical knowledge (basically, I don’t know Ajax, or Flash, or Silverlight, and so if I can figure this stuff out, so can you!).

Over the past two years, at the suggestion of a wise teacher, I’ve peeked at several Edward Tufte books (Envisioning Information, Visual Explanations, The Cognitive Style of PowerPoint, and, Beautiful Evidence), which provide nice frameworks and principles for how we can look at, analyze, and use data. While working at NYU’s Bobst Library, I started to explore Hans Gosling’s Gapminder (which has been featured numerous times on TED Talks, and is based on Trendalyzer software that Google acquired in from the Gapminder Foundation in early 2007). Gapminder allows you to visualize primarily global health and demographic data in bubble graphs over time and by country. Easy-to-use and positively addictive, I think this is a great “entry-level” tool because it makes clear why you would even want to use visual display. It is also a great way to present global health information. Very cool since now I’m assisting Diane Harvey, who is the Librarian for Global Health!

After spending long stretches of time manipulating graphs on Gapminder and making sure that everyone I knew had heard of it, I was primed to see the attraction of data visualization; I started seeing examples all over the place. From Twitter apps to OKCupid’s Trends blog to the Visual Complexity project, data visualization seems to be everywhere these days! Last week, I Googled “data visualization for librarians” and got some really insightful hits to make sense of it all, including lots of work done by Triangle-area librarian, Hilary Davis, Associate Head of Collection Management for Engineering and E-Science Collection Management at NCSU Libraries.

In our last department meeting, Diane Harvey shared her initial experiences and experiments with a data visualization tool called Many Eyes. Many Eyes is an IBM Research project that offers a lot of options for displaying your data. The catch is that you have to agree to have your data public and freely available for anyone else to use (several BYOD -- Bring Your Own Data -- tools seem to have this prerequisite). The visualization options were impressive in scope: Phrase Net, Word Tree, Tag Cloud, Word Cloud, Bar Chart, Histogram, Bubble Chart, Network Diagram, Scatterplot, Pie Chart, Treemap, Line Graph, Stack Graph, Country Maps, US County Map, and some State Maps. So really, lots and lots of choices. Diane was playing around with it just to see what it could do for us given the fact that the libraries collects so much data. Diane’s explorations piqued my interest and inspired me to try, and I’ve been monkeying around, seeing how easy it is to use. Without registering, you can still view other people’s data sets and visualizations. According to the website, you can also create your own visualizations using existing data sets without registering, but each time I tried to do that, my computer froze (a total of five times, and I tried on a PC and a Mac). Registering is easy, though, and things seemed to go a lot more smoothly after I shared my email address. Some exporting is still trickier than I think it should be, but I’ve only just begun experimenting. I’d love your comments or suggestions if you’ve used Many Eyes to present your library data!

These and other examples of visual data presentations (e.g., the Flowing Data blog, Visual Thesaurus, Wordle, graphical tools in Digg) have convinced me that data visualization isn’t just a passing fad; these are tools that I’m going to need to get comfortable and conversant with as I enter the profession. Here’s to a strong start while I’m at Duke! Please let me know if you have any recommendations.

2 comments:

  1. Thanks Alexandra - I've been wanting to explore some data visualization tools. This is really useful.

    ReplyDelete
  2. Awesome, Jean! Thanks for reading. I'm also looking at the open-source Simile Project (out of MIT) at http://www.simile-widgets.org/ since that focuses on visualization as well. There's so much out there, it definitely can get a little intimidating. But I think it's useful to dip our toes in and see what we can figure out together. It'd be great to hear about whatever you find out on your explorations!

    ReplyDelete