Tim Gittos

I'm an Australian currently living in Austin, TX in the USA.

I currently earn a living programming, though I wouldn't call myself a programmer. If I had to attach a label to myself, I'd use the term autodidact.

I love learning, and my favorite things to learn about are programming, computer graphics, AI & machine learning, robotics, painting and creativity.

Using Twitter to Gather Information

Last updated on 30 Jan 2009


I’ve mentioned briefly the virtues I see in Twitter when I discussed how to implement a project in 3 hours. Now I want to expand on how to use Twitter to gather information. I mentioned in that the key to utilising Twitter to gather information was to follow the right kinds of people. However, you’ll find that when you start following more than 100 people, things get a little hectic.

Every programmer is more effective with a solid selection of tools, and Twitter is no different. TweetDeck is the best Twitter desktop client I’ve used, and it’s an AIR app, so it runs on all major operating systems. Most Twitter clients have a single stream that Tweets come in through, and everything gets lumped together. TweetDeck’s strongest feature, for the purposes of leveraging Twitter to gather information, is the option of splitting your stream into groups.

TweetDeck allows you to put people you follow into groups, and have their tweets appear in a different window. The notification alert will also show this segregation. In addition to splitting your followers, you can specify windows that will feed you the results of searches, using the Twitter search operators outlined here.

The search operators and TweetDeck are the engine of our information gathering machine. Using the search operators you can follow the public stream and cherry pick Tweets that contain information you want. You can look for specific phrases, specific hashtags, tweets after a certain date and tweets with links.

For example, I use use a search for “artificial intelligence” with links to find people who tweet interesting links concerning artificial intelligence, which is a subject I’m highly interested in. I can run several of these filters in different windows at the same time, and grok the whole public stream. No longer am I restricted to listening only to those that I follow. Of course, that’s not to say I no longer follow people. If someone pops up multiple times in my filters, I’ll follow them to be sure not to miss a single tweet from them.

This method, however, isn’t flawless. If you look too narrowly, you won’t find any tweets. Also, when you finally do find tweets, even by trying to pick out key phrases/keywords and links, there will be a significant signal to noise ratio. My search for AI topics often lead to students linking university class websites and time tables, which is something I’m obviously not interested in. Often people who pop up in searches all the time turn out to have a high noise to signal ratio.

Having said that, while there are drawbacks, the ability to monitor the public stream as well as the people you follow, and sift out tweets that have a higher likelihood of being relevant is a powerful thing, and will often connect you to bits of information you would have never found otherwise. That’s the power of Twitter, and that’s why I love Twitter.