Community Analysis in general is no brand new issue. The social sciences have been exploring this field of studies for more than 60 years. Various algorithms, which reveal connectivity and activity between people in a social network, have been developed. In addition to that, the graph theory provided some useful tools for efficient analysis. Shortest paths in graphs for example detect optimal positions in a network, from where all members of the network can be reached best (And it doesn't necessarily mean, that these positions have the largest number of direct connections to other members). Finally different methods to express diversity (e.g. color, shapes) have been used regularly.
Social Networking Services
The number of social networking services on the internet increased significantly within the last few years. Communities like MySpace , Friendster or Hi5 host several million members from a lot of different cultures and interest groups. Functionalities offered by these services range from personal profiles, member attributes and friendship lists, over groups, blogs and messaging, to the exchange of multimedia data like music, videos or photos. Despite the differences, nearly all communities share some basic principles. They give their members the possibility to virtually connect, share their interests and extend their social network. But like every kind of new media they have their advantages and disadvantages. On the one hand they extend our sensorium by passing information over thousands of miles to millions of people at the same time. On the other hand over 80% of human communication is non-verbal, so they cut of the main part of our existing sensorium at the same time. They follow their own rules. They have their own dynamics. Here are some thought provoking questions, which would be worth discovering: Do people really want to escape from their social context, or do they rather want to stay in their own town, their school network, their country, their language? Do they break out of our real world borders and interfere with others on an intercultural level, and when they do so, which channels do they use? Can existing trends be detected or even predicted? Why do some interest groups in a given community nearly die out, and why do others remain or grow? Do they falsify information (either on purpose, or by mistake, or by misunderstanding)? Do they shorten our real world activities? These are quiet a bunch of interesting questions and the answers are not trivial. Therefore it is desirable and necessary to retrieve data concerning online social networks, analyze such data, and extract the knowledge we need to answer the questions mentioned above!
Many attempts have been undertaken on the field of community analysis, but unfortunately the most software solutions suffer from a little problem: When the number of displayed nodes increases, their ability to display them properly decreases. The screen seems full and heavy, the user cannot distinguish between important and unimportant information. The chances to improve the existing products and try out new ways of visualization are quiet promising.
2. What is Facelift?
Facelift is a visualization and analysis software for online social networking services. It displays a given community as a node-link diagram and provides several search / filtering functions. The current release provides an implementation for the Online Community Hi5.com. Facelift is divided into two modules: A crawler module and a visualization module. The crawler logs into Hi5 with an existing member account (you have to generate one, in case you don't already have one), downloads member and friends data, and stores the extracted information in a database (by default the Java database HSQLDB is used). Once the information has been stored, the visualization module can display it like you can see in the screenshots below.
Here are some screenshots of Facelift's visualization module.
4. Frequently Asked Questions
What do I need to use this program?
Before you can use Facelift, you need a Hi5 member account. You can get one here.
Do I have do to any configuration?
Yes, unfortunaly you have. The crawler has to be configured in two config files. Here we kindly refer you to the README within the install package and the documentation. The visualization runs without any additional configuration.
How does the visualization work?
The visualization is build on top of a visualization framework called Prefuse, written by Jeffrey Heer. At this point, we would like to thank Jeff for his preliminary work. Thanks a lot.
Can I join the project?
Of course you can! We are always happy to welcome new developers.
Just visit our Sourceforge Project Side or contact us directly.