Web Retrieval and Mining - Final Project

Interactive Social Network Visualization - FriendViz

R96944043 周伯相 R97944018 李庭嫣

http://www.cmlab.csie.ntu.edu.tw/~zho/

Final Project

We constructed an ego-centric view of the social network by mining users' friend list on the most popular social website, Facebook. With layout placement policy and interaction tools, we help users identify the cluster of their social network and the relationship among their friends.

Implementation

• Cluster Identify
  With our work, it can be directly found by viewing the graph. Also, the visualization helps us to find the mediator among groups.
  

• Close View
  Besides knowing an overview of whole friends, dragging the friend node helps us see more detail relationship among those friends.
  

• Highlighting, 3D Traversing
  The 3D visualization helps user doing randomly rotation to see the structure of the social network. Moreover, when user clicks a friend, the graph then moves the specified friend to the center of the graph, and the link of the friend will be drawn in yellow to help users find the friends of the specified friend.
  

Result

• Gathering the info of friendship
  We parse the web pages on Facebook to obtain the relationship between friends.

• Node Placement Policy
  We use force-direct layout to place the nodes by assigning spring forces and friction on each link.
  

• Avoiding Overlap
  To avoid the overlap, we applied collision detection in screen space (2D space).

• 3D Acceleration
  In order to display dynamically massive nodes and links, we use hardware acceleration, Direct 3D 9. By the modern hardware, the limitation of node display is more than 1000 nodes.

Download

Project program: Vis3D.rar

Video: vis3d_vcd.mpg

Reference

• Facebook
• TouchGraph Photos
• Wasserman, S., and Faust, K., 1994, Social Network Analysis: Methods and Applications (Cambridge, UK: Cambridge University Press).
• WIKIPEDIA, Force-based algorithms, http://en.wikipedia.org/wiki/Force-based_algorithms
• WIKIPEDIA, Information visualization, http://en.wikipedia.org/wiki/Information_visualization