I worked with irhetoric on many applications, one of which is The Archivist – a WPF Windows Application used to archive and export searches on Twitter. He and I have been both using The Archivist to save Tweets about Iran, and the recent problems around the election there.  As the use of Twitter has been advertised so heavily in the news, we thought it an interesting focus for this experiement, as we could “capture history,” as well as pull an amazing dataset of information that could be of great importance.  So far, we each have over 200,000 Tweets saved in xml files.

With so much information, what could we do with it?  I asked irhetoric what the possibilities were regarding how much information we could pull from this data: location? Username?  Friends?  Almost anything is possible.

Mix Online published the Descry Prototype back in January of this year.  The project centered around data-visualization and, specifically, infographics.  I thought that I might be able to contribute to this project by building some myself.

I started in Photoshop, and produced some comps:

In this first example, the idea behind increasing the size of the avatars of the users who are “most-followed” is to be able to, at a glance, see who the most “influential” people are. If someone is highly followed, it stands to reason they reach the broadest audience, therefore making sense to reach out to those users first.

Placing these avatars on a timeline, as well, you can see the more active parts of the day. We’re also considering the potential of adding a third vertical axis that would deliniate the number of Tweets a user has sent.  If someone Tweets a lot, but is not very influential, that would be different than someone who was very influential and Tweets a lot. Also, someone who is very influential but doesn’t Tweet much, might just be a celebrity or high-profile user. Again, this user has very different value than one who provides lots of good information, and is very influential.

Now, all of those numbers above are totally made up … lorem ipsum, if you will.  The idea, though, of placing the “hotspots” from where people are Tweeting, I thought was interesting, especially if we bind that to time, so we could see the influence of Twitter, and how it affects global communication.

I designed this bar graph to demonstrate the volume of Tweets over time.  The pink graph represents the Twitter-search: “Iran” and the blue bars represent the Twitter-search: “northkorea”.  Again, all data is fake here, and that’s why I needed irhetoric to help me with the data.  

Irhetoric was able to take some of the code that our MIX Online team member, Hans Hugli, wrote for Descry. He took his Tweets and created a basic tree-map visualization based on the number of Tweets associated with a user.  He took the top 100 Twitter users, and displayed their avatars, usernames, and number of Tweets.

I'm looking forward to developing these ideas, and hope that anyone interested in implementing these data-visualizations should absolutely feel free. We hope you'll let us know were we can find your project. Be sure to download The Archivist to download, save, and export Tweets you care about. Leave a comment below and follow us on Twitter @MIXOnline!