Cite: P. Appan, B. Shevade, H. Sundaram. Interactive Visualization and Content Analysis of Instant Messaging Networks, in Arts Media and Engineering Program, ASU, AME-TR-2004-14. December, 2004.
Abstract: In this paper, we present our efforts towards allowing users to query and reflect upon their instant messaging histories. The problem is important since the volume of chat data is increasing at a significant rate, and traditional text based retrieval does not capture the social conventions in such systems, allowing no space for insightful exploration. In this work, we develop an analysis and interactive visualization framework, to provide insight to the user. In addition to the traditional text retrieval features like dominant concepts and their frequency, we propose to extract novel features specific to instant messaging domain like, usage of emoticons, use of exaggerations and emphasis words, number of foreign language words. These features are used to determine the key aspects of communication (a) between two users, (b) between a user and different categories/groups of people like friends, family, colleagues etc. Finally we develop two novel exploratory schemes to explore an archive of a group of users to help the users reflect on the nature of the communication (concepts, network temporal evolution, and communication density).
[ download ]