« « Data Delver: Lisa Pickoff-White, California Watch

Data Delver: MaryJo Webster, Pioneer Press » »

Many Eyes: “Catalyzing the community around data”

Posted by on Feb 18, 2010 in Blog, data visualizations | No Comments

Before I get going, thanks to all of you who’ve been reading and/or sending me feedback!  And a hearty welcome to any new readers who’ve found their way over via Innovative Interactivity, #wjchat or any other method.  I’m so glad you’re joining me on my data journalism quest (read: obsession!).  If you have any topics you’d like to see explored, or tips and advice, just let me know.  Okay, seats buckled?  Off we go!

I clearly remember the moment I first realized the power of visualization for reporting, as opposed to presentation.  It was a typical Thursday evening in Medill’s DC program.  Every week, for three hours, we got to participate in data nirvana — more formally known as Derek Willis‘ CAR class.  That night, he showed us a tool called Many Eyes — it’s visualization for the rest of us.

Looking at spreadsheets is great, and I find it fascinating.  Many of us got into CAR, I imagine, because we like to find the story in the numbers.  And scrolling and resorting is a fine way to start.  But if a picture can do it for us, and there’s aesthetically pleasing colors, that’s even better, and we may discover even more.  But visualization can be complicated and time-consuming, not to mention the knowledge of programming and design it demands.  How can we do it quickly on deadline to inform either ourselves or our readers?  Enter Many Eyes.  Input your data set, numerical or textual, (but make sure you are okay with it being public info), and then visualize it in all sorts of ways.

Diving into text viz has been a more recent foray for Many Eyes.  The main part of text viz as I previously understood it means the tag clouds and wordles you’re used to.  I frankly agree with Mark Schaver that they “belabor the obvious.”  I also think they are often overdone and aren’t that useful.  Why?  They lack context. Sure, McCain mentioned Obama a lot in a campaign speech, but what connotation did it have? What about Obama was he discussing?  But there are better ways to do it, and I know there are interesting stories within text.  If anyone knew the answer, it would be someone from Many Eyes.

That’s why I jumped at the chance to attend a talk today when I heard that co-founder of Many Eyes, Fernanda Viegas, was coming to Northwestern’s campus today to discuss viz for both numbers and text. Here are some takeaways:

  • This is an exciting time for visualization, as it’s moved from the scientists to the artists to the journalists to everyday folk. Viegas said ManyEyes was created so you don’t have to cross as many barriers to create a visualization – programming, interface design, etc.  She’s not concerned about people using visualizations improperly due to lack of training, but believes that rather, as more and more people play with ManyEyes, their data literacy will increase.
  • People see visualizations as a personal way to express themselves. In graduate school, she worked on a project visualizing participant’s email archives.  She paid excruciatingly detailed attention to  privacy concerns.   But when participants saw the visualizations, the first they wanted to do was share them with others.  “They thought of the pieces like photo albums,” Viegas said.  “They wanted to look at them with others and reminisce.”
  • Viegas used journalism as an example of showing how far we’d come in the viz world.  She said visualizing the Billboard ranking of Michael Jackson’s songs as a way of celebrating his life, as the New York Times did, would not have been considered at the beginning of her career.
  • There’s a strong user-generated content and community component to Many Eyes.  As people look at each others’ data, they find anomalies that the original uploader might have missed, and they chat with each other about it in the comments, showing the visualization at the frame they interacted with it to make a point. By doing this, I think they’ve made an important breakthrough by allowing to essentially bookmark a frame of an interactive as you saw it.  I don’t think I’ve seen that elsewhere, and I know it’s one of my issues with Flash.  (In journalism, I’d love to see more of people making discoveries like this from visualizations, and even data centers, they might find on a news Web site.)
  • A lot of people are visualizing data you’d expect, aid to Haiti, budget numbers, the King James Bible.  But people are using visualizations in their personal lives.  Two favorite examples was a treemap sorting through a wedding guest list or the text of personal ads on Craigslist.
  • To give context to text visualizations, you can use a word map, like the Bible example above, which shows which words come before and after a word you input.  Phrase Net is the newest text viz option.  For example, it will analyze text and relate words “A” and “B” every time it sees the phrase “A and B.”  So, in Viegas’ example of the text of Pride and Prejudice, you can see which characters are mentioned in the same phrase, and how often. With a click, you change the visualization, so it uses “or” as the key for connecting two words. There are other options as well.  The best way to experience this is to check it out for yourself.
  • “We are used to thinking about visualization as a neutral tool expressing hard facts,” said Viegas.  “But it really depends on what you decide to represent, and how you decide to represent it, you have a lot of say in the message you convey.”
  • She pointed out that people even have political debates with visualizations.  Someone posted a word tree demonstrating how often Alberto Gonzales said, “I don’t recall” in his Congressional testimony.  It was featured on the front page of Many Eyes, and someone else responded with a similar visualization based off the text from former President Bill Clinton’s testimony about the Monica Lewinsky incident, pointing out that there was also a lot that he didn’t recall.
  • All Many Eyes visualizations are embeddable and shareable, to capitalize on what Viegas learned in that first project.  “It’s much easier to talk about a picture than a table of numbers,” said Viegas.  “We’re helping catalyze the community around data.”

The bottom line is that there’s a lot to explore here that we can use to help us find stories in data, or can present to readers to either illustrate a point or allow them to make their own discoveries.  I should point out that all Many Eyes visualizations are embeddable, so can be used in stories or blogs.

And just for fun, although not particularly journalistic, here’s a tool that Viegas and co-IBM research scientist and Many Eyes co-founder Martin Wattenberg created that allows you to compare what people are searching for, using Google Suggest as data.  Viegas did some fun comparisons “Why doesn’t he” vs. “Why doesn’t she” or “Are dreams” vs. “Are nightmares” or “Is my son” vs. “Is my daughter”.  It’s a fun toy, and gives some good insights into crowd psychology. Or shows, as Viegas put it, “People think Google is like an oracle, that it can solve their problems.”

« « Data Delver: Lisa Pickoff-White, California Watch

Data Delver: MaryJo Webster, Pioneer Press » »