In 2007 IBM launched the site Many Eyes, which allows users to upload data sets, try out various ways of visualizing them, and most importantly, discuss those visualizations with anyone who sets up a (free) account on Many Eyes. As professor Ben Shneiderman says, paraphrased in the New York Times review of Many Eyes, “sites like Many Eyes are helping to democratize the tools of visualization.” Instead of leaving visualizations to highly trained academians, anyone can make then and discuss them on Many Eyes, which is a pretty neat idea.
Many Eyes allows viewers to upload data sets and then create visualizations of them. Many Eyes offers users the ability to visualize data in 17 different ways, ranging from the wordle type of word cloud, to maps, pie charts, bubble graphs, and network diagrams, just to name a few. There are other sites or programs that will allow users to create charts in some of these ways, Microsoft Excel for example, but Many Eyes offers the advantage of multiple types of visualizations all in one place.
Additionally, people in disparate locations can talk about the data sets and visualizations through comments. The comment feature even allows for the “highlighting” of the specific portion of a visualization you might be referencing. The coolest feature of Many Eyes is that anyone can access and play with data uploaded by anyone else, in the hopes that “new eyes” will lead to surprising and unexpected conclusions regarding that data.
If you create an account on Many Eyes, you can access their list of “Topic Centers”, where people who are interested in data sets and visualizations relating to specific topics, can interact and comment with one another, as well as link related data sets and visualizations. However, a quick perusal of the topic centers show that the vast majority of topics are being followed by only one user. The few topics that have more than one user seem to be pre-established groups with specific projects in mind.
Unfortunately, it appears that a crowdsourcing mentality, where people who don’t know each other collaborate to understand and interpret data, hasn’t really materialized. In this IBM research article, the authors even hint at how Many Eyes “is not so much an online community as a ‘community component’ which users insert into pre-existing online social systems.” Part of the difficulty in realizing the democratizing aspect of Many Eyes might be a simple design problem in that the data sets, visualizations, and topic centers display based on what was most recently created, rather than by what is most frequently tagged or talked about. This clutters the results with posts in other languages or tests that aren’t interesting to a broader audience. Many Eyes developers might adopt a more curatorial method where they link to their top picks for the day on the front page in order to sponsor interest in certain universal topics. But maybe the problem might be more profound; what do you think?
Ultimately, I’m not sure how relevant Many Eyes is to historians. It seems that asking for a democratized collection of strangers to collaborate on visualizing your data seems unlikely based on the usage history of the site. However, groups of researchers who already have a data set to visualize and discuss might be able to make use of this site for cliometrics-style research. Classrooms and course projects in particular can benefit from this site, since it’s relatively easy for people with a low-skill level to use. What do you think? What other applications do you see Many Eyes having? How relevant will it be for your work in the digital humanities?
