If You Can Analyze a Bunch of Tweets…

If you can analyze a bunch of tweets, what would you like to find?

I found a utility to back up my tweets called bmtu. It is a pretty cool python program and I have been using it for a while. If you are not into python and would like to get a Windows executable, send me an email and I will be happy to send it to you.

This program (bmtu) not only lets me back up my tweets but also save Tweet Streams of other people. So I picked a few influencers in the industry and saved their tweets as well.  Each tweetstream goes into an sqlite database. I have been doing this for a few months.

I have a few ideas on analyzing these tweets. I may build a few of these (if I can’t find other programs to do this).

  1. I would like to take my twitter stream, extract all the links and stick them in a social bookmarking site like delicious.com (I use Twitter as a bookmarking service. However, since it is not easy to locate tweets beyond three days, I would rather have them on a social bookmarking site where they can be shared and tagged by any one).
  2. I would like to parse the text in the tweets, remove the common words and create a vocabulary (this may require a couple of iterations to remove noise but in the end should give me a nice vocabulary of my tweet-o-sphere).
  3. Take 2 one step further and create a tag cloud and link them to the tweets (we have already built some kind of linked tag clouds and find them pretty useful).
  4. I would like to repeat 1-3 with the tweets of some of the people I follow – you can guess where I am going with this.
  5. I would like to create an ontology of topics in my tweets. It is easier said than done. Even though the starting point for this comes from 2 and 3, I need to process them a bit more and consolidate them.
  6. I would like to find out which topics get more retweets than others. This will give me a sense of what my readership likes.
  7. I would like to find out my own retweet patterns. What do I retweet most? Who do I retweet? What topics do I tend to favor? Are these the same topics that I tweet about?
  8. I would like to find some propagation patterns. For example how many people who retweet my tweets are not directly following me?
  9. What percentage of my tweets get retweeted? What percentage of links I post, get read?
  10. I follow some thought leaders like Tim O’Reilly who tracks alpha-geeks. I want to see a timeline of his tweet topics and get a sense of where he thinks the industry is moving.

What about you? What would you like to know about your tweets and the ones you follow?