• Print

Twitter Approval Matrix

There is a lot of interest in figuring out how to most effectively use Twitter. The Twitter Book is a great start to put some of the of the puzzle pieces together. The book got me thinking about some of the people and tags that are popular. I was amazed at how Ashton Kutcher, @aplusk, surpassed 2 million followers relatively quickly. My curiosity led me to scraping his tweet archive and producing a Wordle of his tweets. I found very little substantive content as compared to Kathy Sierra (@kathysierra) or Tim O’Reilly (@timoreilly). Don’t get me wrong, Ashton does tweet, but it’s about boring stuff, IMO. But boring to me could be exciting to one of his followers.

This got me thinking further about tastes. We all have different tastes. I like to make sense of the world around me by collecting as much information as I can, analyzing it (in my head and in programs), and then visualizing it. So, that led me to the matrix idea. Why not plot Twitter activity, trends, users, tags on a matrix? For years, I’ve been reading the New York Magazine and always scanned its Approval Matrix. They describe their Twitter version as, “Our deliberately oversimplified guide to whose tweets are worth following.” I thought I’d do the same, but with more than just who but also what is worth following. I decided to throw in some analysis and a user contributed component: you. The “you” part is for the future. That is, I would like you to contribute coordinates for where you think items should land on a future grid similar to the one below.

This matrix shows four quadrants used to describe tastes found on Twitter, or related sites such as hashtag.org, tweestats.com, etc. The Y-axis is partly analytical and shows popularity (mostly through scraped numbers) or perceived popularity (in the future nominated by you). The other part of the grid is more curated and subjective. The X-axis has been plotted based on my personal opinion. You may agree or disagree with my placements and that’s all good to me. After all, it is about taste. The matrix and plots do not represent a thorough analytical treatment, but rather a view of the trends that could be found in data sources allowing me to plot with some sense of relevance.

TwitterApprovalMatrix.png

For this post, I’ve limited the data and activity to the month of May. I will make this a monthly post if I get enough feedback/help. So, here’s how you can nominate topics or people for the matrix:

  • You can tag any of your tweets, RTs using #approvalmatrix and I’ll find it.
  • Tweet to @mikehatora
  • DM mikehatora
  • Send email to mikeh {at} oreilly {dot} com

If you want to suggest where the tweet or subject belongs on the matrix, do this:

  • Notice the numbers on the grid, 1-10, in North/South and East/West directions.
  • Notice the quadrant tags NE, SE,NW,SW, on the outer corners of each quadrant.

So NE, 6,8 would put your suggestion between #jobs and #followfriday.

Here are a couple of examples that could end up on future ApprovalMatrix postings:

  • @mikehatora #wwdc, NE 5,9 Interesting new stuff from Apple, but seems to have NOT leapfrogged ahead again this year. The plot is hot [N], because lots of hastags have #wwdc and the comment shows it is interesting [E] but not at the 8-10 scale of smart but rather a 5-right, 9-up on the plot]. You could say a topic like this is hotter than it is smart/interesting/useful.
  • Starting new job on Monday, thanks to Twitter posting. #ApprovalMatrix This will be picked up in a scrape and added to the jobs/careers bucket. It’ll likely will be a NE 5,8 aggregated with the other job data and comments.
  • #approvalmatrix NE 7,8 RT @timoreilly Three-part series on how Google does search quality evaluation, starts here: http://bit.ly/QIgwX Here the tweeter is Re-Tweeting someone (timoreilly) and indicating that it is an important item in the Hot and Smart quadrant (NE). I’d look to see how many RTs have happened after the initial tweet.
  • DM mikehatora SE 7,8 #hr Indicates the hashtag #hr, which is Human Resources, is not trending up, and is boring according to the sender.

Feel free to nominate people, tags, related news, etc. I can’t guarantee your nomination will make it into next month’s post, but it just might. This is a chance for you to put things on our Radar while letting us know what your tastes are.

I hope you enjoy this and see it as a potentially useful tool to monitor trends that your fellow readers are tracking.

tags: , , , , ,
  • http://friendfeed.com/bjmendelson Brandon Mendelson

    How are you determining who is and who is not a celebrity? I count big brand news agencies as well as web celebrities, so I think it’s far less than 77.

  • http://twitteroftheday.com Noah

    I do believe isolating Twitter signal from noise takes, like you say, a mix of analytical metric, some personal opinion, and the voice of the crowd.

    I created a fun site called http://TwitterOfTheDay.com that uses those same ideas. People can nominate their favorite tweets easily (although we also have some organized site contributors) and anyone can vote on whatever they like.

    It’s Digg for Twitter, helping people find the most valuable users of Twitter and celebrating those that make it a useful place to hang out!

    However, your consideration of topics of interest is great. I’m curious what other sites & apps will come up to help people find who to follow!

  • http://www.twitter.com/Open_Book Alistair Spalding

    You’re going to need a bigger graph!

    I’d be more interested in something that seeks out and promotes the less popular but very interesting areas of twitter. Arguably the things shown on the graph are the obvious/well known/popular places.

    I think if it began to really map out the extent of everything people tweet about it would be very special.

  • http://andrew.blumlein.net/index_gb.html Andrew Levine

    Easier than “NE 6,8″ would be “Hot+6 Smart+8″, with the axes going from -10 (Boring / Cold) to +10 (Hot / Smart). With the points of compass one has to consult the matrix before placing a tweet.

  • http://www.chatlobi.com Seviyeli Sohbet

    Cıvıltı için Digg’dir, cıvıltının en değerli kullanıcılarını bulmak için insanlara yardım etmek, ve dışarı asması için onu faydalı bir yere yapanları kutlamak!

  • http://twitter.com/GregJaros GregJaros

    Nice job here. Good visual. FYI there is a spelling error, “Chiacgo”.

    You have aplusk in the right spot – surprisingly boring for someone with his lifestyle.