Monday, August 22, 2011

Viz: Twitter Network Exploration

Mentionmapp- An interesting web app for exploring twitter connections is a good example of a data visualization. This app basically asks the user to input a twitter user name and displays all the other twitter users who interact the most with the given twitter username, and also the hashtags mentioned the most by the user.

The twitter user name could be a celebrity or even a friend’s account. The username accounts that can be searched through the app must be a public profile, that is the user must have allowed public access to his account.
The visualisation runs in the browser displaying the relevant results. The system works as follows :
As soon as we enter the twitter username to be followed, the app API, loads all the tweets associated with the user, computes the people and the hashtags interacted with or mentioned the most by the twitter user and displays them in a visual format.
The data is developed using constellation framework. Connections between users reveal interactivity levels and discussions aggregate into clusters.Details about the twitter user’s accounts are revealed in a panel on the left.These details include id,name,location,profile and URL of the twitter user’s account.
Each user’s profile picture is also displayed,along with the username and connections to the most mentioned hashtags.Hence a graph representation of different twitter accounts with nodes representing users and edges representing interactions is presented. The edges between the nodes are bigger if the number of interactions between the two sets of usernames are larger.This draws users towards bigger discussions.Hovering over an edge reveals the exact number of mentions.
Clicking on a specific node ,enables mouse over effects,to emphasize on sub networks.
There is a zoom in/zoom out panel on the right.The top panel also displays a history of searches .We can go back to a particular username searched earlier by clicking on the username in this “history” top bar.
The data from Twitter doesn't come all at once and the visualization reflects that. As soon as new information is loaded, nodes and edges will appear and the layout will shift to accommodate them.
Other drawbacks include inconsistent performance due to the twitter server being temperamental in its response.The failure cases are fairly handled with messages like “server unavailable“ and “over capacity” displayed.
The URL for this app is :
Reference :