Visualising your Twitter conversations

2009 November 9
by velvet

Several weeks ago, I came across a tweet about visualising Twitter conversations. Curious, I checked out the article and what I found was simply fascinating, not to mention intriguing works of art. Cate Huston (a.k.a. ‘Kitten the Bad‘) is a Masters student in Computer Science at the University of Ottawa. Cate devoted an article on her blog to mapping peoples Twitter conversations. The results are striking. By adding levels of conversations, she could create a graph that showed the Tweep’s conversations with others, and, in turn, their conversations with other Tweeps.

Cate’s rationale for creating these graphs is to differentiate various types of Tweeps. Some are conversationalists (the networkers), while others are broadcasters or spammers. If a graph is generated for multiple levels of conversations, it can also show which groups (or sub-networks) the Tweep is most active in and the Tweep’s level of influence in different networks.

I asked Cate to create graphs for my two Twitter accounts, Velvet Escape and Velvet Connect. The results are shown below.


Velvet Escape talks to A LOT of people, and it seems, Velvet Escape is talked about a lot, so Cate created a graph that only showed the Tweeps Velvet Escape engages with directly, and those who talk about Velvet Escape (single-level conversations). The result was stunning. The yellow lines represent a reciprocal relationship while the red and purple lines represent a one-way relationship. The graph shows that Velvet Escape has many reciprocal relationships. These reciprocal relationships in turn signify the level of engagement with followers: interaction and re-tweets (how much a person re-tweets and is re-tweeted).

For Velvet Connect, Cate created a graph with multiple levels of conversation. The result is strikingly different. Velvet Connect’s sub-networks can be clearly seen, but, probably as a consequence of Velvet Connect’s relatively new presence on Twitter, the number of reciprocal relationships (compared to Velvet Escape above) is smaller. However, the community-building efforts are quite evident. Reciprocal relationships have been established with various key networkers – see the yellow lines to the larger sub-networks, especially those in the middle and top-half of the graph. I suspect Velvet Escape’s graph would have looked pretty much the same six months ago.

velvetconnect graphThese graphs say quite a bit about a person’s behaviour on Twitter. Is the person just shooting out tweets with links or is the person engaging with his/her followers and active in conversations? Mapping someone’s Twitter conversations certainly is a great way to visualise, and hence measure that person’s level of engagement.

Great work Cate!

Check out Cate’s Twitter graphs.



No comments yet

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS

Get Adobe Flash player