http://www.cs.kent.ac.uk/pubs/2004/1931/content.pdf

Download PieSpy: http://www.jibble.org/piespy/

PieSpy is a freeware program that can be used to track interactions amongst users in an online chat room.  The software employs several methods to not only determine which users are interacting with each other, but also just how strong these interactions between the two individuals are.  It then generates a detailed mathematical model for the Chat Room Social Network.  The program uses three methods to achieve this.  First, it stores each individuals nickname in the chatroom, and when a chat post directly addresses someone, the two individuals participating in it are linked.  Secondly, the software uses the concept of temporal proximity.  If no message has been posted in several moments, and then two are posted close together, it is reasonable to assume that the second message was in relation to the first and thus the two individuals are connected.  Finally, the program employs the notion of “temporal density” when it analyzes messages.  What these means is that if n messages have been sent over a short period of time and only two users are involved in the sending, these two users are assumed to be having a conversation.  The program uses a setting of n > 5 for its “conversation” threshold.  The length of the line connecting two individuals is proportional to the strength of their ties, which shorter lines connecting those that are “stronger” in their linkage.

By tracking and evaluating the interactions between people in a chat room, PieSpy enables the user to construct elaborate social networks for groups of people.  Furthermore, by labeling the strengths of each edge, it allows some approximation of just who has power in a social network.  As discussed in lecture, those with a high centrality to a network or multiple strong ties exhibit more power than others in that network.  The program allows easy visualization of this concept.  Furthermore, each time PieSpy is run in the same IRC channel it bases its network off networks that have already been stored. This allows one to see just how the social network has evolved over time, providing insight into just how power shifts in an ever changing social network.