How Can We Study The Data of Social Networking Sites?

On: October 6, 2008
Print Friendly, PDF & Email
About Bram Nijhof
Bram Nijhof is a master student of New Media at the University of Amsterdam. He has a bachelor degree in Art & Technology and a bachelor degree in Media & Culture with an expertise in New Media.

Website
http://www.bramnijhof.nl    

How can we make use of all the data stored in Social Networking Sites? And how can we study those data? In 2005 Jeffrey Heer and danah boyd developed a system called Vizster. It is an information visualization system that visualizes networks within the Friendster social networking service. Within networks there can be seen how networks are related to each other and there can be searched for example for keywords or gender. Data can be visualized in different ways and they are shown in another form then how the Friendster service is displayed. New contexts will be created when visualizing the relationships or the differences between people, interests, background or gender.

The Digital Methods Initiative has done projects on Social Networking Sites. They call the use of the personal data in Social Networking Sites ‘ post-demographics’. Not only traditional data such as age, gender or income can be used, but also interests such as music, books or movies. An example question they asked is whether friends have the same interests or not.

DMI has some examples of systems that make use of the data visualization. One example is Elfriendo.com. With the tool one single interest can be related with other interests those persons have. Or the interests of the friends of a politician can be compared with another politician. It is possible to see which movies or music the friends of a politician like. All data is scraped from their profiles and this is done for Myspace. Another example is the Dutch tool Friends’ Politics (Vriendjespolitiek.net). The tool uses the Dutch Social Networking Sites Hyves to tell people what their characteristics and their friends tell about them. It gives an overview of which politician fits the best into their profile. For example different universities in Hyves differ from which politicians are most popular. So some universities or companies can be more left- or right wing. If you have a company or university in your profile then it has influence on the politician who fits the most into your profile.

The Vizster system was mend for end-users to let them discover and make them aware of their online community. With the tools described above it is possible to let users see what can be done with their information. Heer and boyd said that is was more for play than analysis. But are there possibilities now to do serious research with this data? The Digital Methods Initiative has put some questions on their Wiki. They made a distinction between the demographics as age, income and location and the post-demographics where other interests out of the Social Networking Sites are important. They asks themselves the question: When, and for which purposes, are interests a more significant mode of organizing, sorting and recommending action than demographics? This is another way of studying groups in stead of the traditional way.

Last year in the DMI I proposed a ‘Localizing Hyves’ project to map the friends of politicians into Google maps. You can see where the friends of a politician are living. You also can make a tool where you can see where certain groups or groups with certain interests are living. I think with all these amounts of data in the near future there will be new tools developed. But I think it is a challenge in analyzing and interpreting the data.

There are more examples than the above ones. Last year there were made some interesting projects in the course ‘information visualization’. Students from different disciplines have made applications which are able to visualize networks. In the posting Visualizing the network
these projects are presented.

Jeffrey Heer and danah boyd (2005). “Vizster: Visualizing Online Social Networks.” IEEE Symposium on Information Visualization (InfoVis 2005). Minneapolis, Minnesota, October 23-25.

Leave a Reply