Adding context to social media through tech


 

Like a lot of people, I keep a social-media dashboard on my computer screen all day, usually in the background. Some people listen to Pandora while they work; I “listen” to the hundreds of people in my various social media feeds.

I come by this naturally. My grandmother, like most of my family, lived in a small, Southern Indiana town. She kept tabs on the community with a Cobra police scanner, which enjoyed a prominent spot in her living room. Visits there – including Thanksgiving dinner and gift opening at Christmas – were accompanied by local police, sheriff and volunteer fire chatter.

Listening to social media is like that to me: an ongoing river of updates. As a consultant, it’s how I keep track of multiple clients and their interests, keeping half an ear tuned for the occasional emergency.

But while my dashboard (I use TweetDeck; my wife uses HootSuite; vive le differénce) keeps me tuned in to the ongoing chatter, its immediacy and universality lack something important: Context.

If you follow someone long enough, you’ll know their opinions and interests. You’ll get a gut feel for that, the same way you’ll get it after talking to someone for an hour at a cocktail party: Bob likes the Reds, his kids are academically exceptional but empathically dim, he rides a Harley, his wife runs her own company, he’s a progressive and gets a headache from red wine.

All very good background. The problem: Each one is equally weighted. In a river of chatter, you only have your gut to tell you what Bob really cares about.

That’s why I was intrigued this weekend at the annual Block by Block conference (which The Patterson Foundation helped to found and has been the key supporter of for three years) to see a new development by the Knight News Innovation Lab at Northwestern University. It’s called TwxRay (pronounce it TwixRay and you’ll save yourself a tongue-tying), and it brings context to Twitter’s chaotic chatter.

The web-based tool allows you to analyze a Twitter user’s activity, and then see a visual pie chart of what are the most common themes that user discusses via Twitter. That way, you can see the top topics discussed by the Twitter user, weighted to how often she tweets on the topic.

For example, by applying it to @ThePattersonFdn, you’ll found that our most popular topics, by percentage of tweets in those topics, are:

  • Impact:                        29%
  • Education:                        25%
  • Healthy Living:            12%
  • Technology:                        7%

And for our CEO, Debra Jacobs @debramjacobs:

  • Education:                        40%
  • Impact:                        18%
  • Technology:                        6%
  • Healthy Living:            6%

And so on. It analyzes the last few hundred tweets and categorizes them using a sentiment-analysis back-end engine called Alchemy. Since the technology is statistical in nature, it will sometimes make mistakes – but it sure is fun to use.

At TPF, we also like thinking about this tool as part of our partner analysis – listening to a potential partner’s activity in social media to help determine fit long before you sit down to your first brainstorming meeting.

The Knight News Innovation Lab, housed at Northwestern University, is charged with developing tools for news and community engagement. Some of them are very useful, some not; some of them are ready for outside use, some not.

TwxRay is still in development, and its tagging categories could be more nuanced, but, like my grandmother’s scanner, TwxRay now has a spot on my digital mantle.

 

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>