Visualization and Learning Analytics

This project seeks to develop an interactive visualization system to investigate group and individual differences in game play strategies and decision making over time.

So far the group has developed a visualization system called Glyph that shows paths through game states and how many users are going through them. Interactivity is used to allow users to isolate specific traces and see how many users took them. Thus, allowing designers to see dominant or popular strategies and not so popular ones. They can also clearly see winning vs. loosing strategies. More work is currently underway to look into separating the traces by different rewards to see individual differences and preferences.

behavior_graphstate_graph

This  graph shows nodes of all paths, the bigger the paths the more popular it is. The number against the node shows how many users used this path. The color shows if the path led to a dead end or not; red means there is a dead end.

 

 

This graph shows all paths that players took. Nodes are states in the game and the links are transitions between states. The red nodes are end states. The blue node is the start state.

 

 

 

 

Doctoral Fellows: Shree Durga, Truong Huy Nguyen Dinh
Faculty Members: Alessandro Canossa (PI), and Magy Seif El-Nasr (Co-PI).

The project was funded by Northeastern University Interdisciplinary Tier 1 grant.