Facebook’s Lexicon dives deeper to show you what’s hot with Facebook users

While the rest of Facebook stresses about negative user reactions to its new redesign, a team of its developers has launched an exhausting new version of its popularity-analytics service, Lexicon.

Lexicon’s new features dig deep into the company’s 100-million-user-strong data set. The original version of Lexicon launched this past April and only showed you the frequency of various words as they appeared on users “wall” message boards across the site. The ‘frequency’ was derived from comparing mentions of a given word versus mentions of all other words during a given period of time.

Today, a new “dashboard” tab shows you the total number of unique users who mentioned a topic, the percent of the user base that mentioned the topic, and the total number of posts on the topic. A new “demographics” tab includes ways to track trends over time based on age, gender, and country. A new “associations” tab lets you see relationships between a topic keyword and related words over time. A new “sentiment” tab shows you whether most people expressed positive or negative feelings about a topic. A new “pulse” tab shows you which Wall topics also show up within user profiles, such as ‘favorite music.’ Finally, a new “maps” tab maps out topical data and comparisons between words by state and country.

Exhausted by the possibilities?

Well, just try digging in to get the data to come to life. For example, I’ve compared mentions of U.S. presidential candidates John McCain and Barack Obama across the 50 U.S. states over the past week, using the “maps” feature. (See screenshot, above.) Obama has generally gained more support from Facebook’s core demographic of college students and 20-somethings. But, by using the “relative comparison” option within this mapping, you can see that McCain is getting talked about a lot more than he used to be — except in some key Rocky Mountain battleground states.

It’s not clear what exactly the relationship is between presidential polls and what Facebook users are talking about. If I were in either campaign, though, now would be a good time to start playing around more with Lexicon. Understanding Facebook users’ feelings about the respective candidates is strategic information for figuring out where candidates should campaign.

Meanwhile, Lexicon is still a work in progress, and Facebook only offers sample topics to look at for now. But it plans to offer more topics, and support additional demographic and country information in the coming weeks. One use of Lexicon that I’d love to see: What are Facebook users saying about Facebook’s redesign.

Next Story: Massive iPhone 3G power adapter recall for U.S., Canada, Japan and Latin America
Previous Story: “I’m a PC” and I’m made on a Mac?

Bookmark and Share

Tags:

Photo of Eric Eldon

About the Author, Eric Eldon

Eric currently covers digital media technology and business news, especially what's happening on social networks and their platforms. He also writes and edits stories about venture capital, and lots of other stuff, too. He started at VentureBeat in the spring of 2007, half a year or so after Matt Marshall left his reporting job at the San Jose Mercury News to found the site. Eric previously cofounded a startup called Writewith, that was building editorial software for newspapers and other groups of writers. The startup didn't work out, but he learned a lot.

  • The new Lexicon app is 'pretty cool' but incredibly restrictive - especially since you can only choose from 17 topics to track.

    The really value in this tool will be when any word can be tracked and displayed long with sentiment. As an interactive & brand strategist i'm excited by the possibility of monitoring mentions of my clients branded keywords in facebook and seeing a) how they trend and b) what the sentiment around them is.

    Once Facebook ads a last layer to them mix - integration with their social ads platform - we'll be able to see in near real time the effect of social ads on sentiment around brands.
  • Peter, I agree. I wonder if letting third parties track custom data is going to be a premium service? Might be an interesting secondary revenue model.