Data extraction from social network applications
Nowadays, there are several social network applications available on the web. Some of these social network has a specific propose to gathering people, others are just general propose and competing for the user attention. For example, my favorite social applications are Facebook (helps you connect and share with the people in your life), Orkut (Connect with friends and family, Discover new people through friends of friends and communities, Share your videos, pictures, and passions all in one place), and LinkedIn (Stay informed about your contacts and industry, Find the people & knowledge you need to achieve your goals, Control your professional identity online). In these systems, people share information freely, with restriction about privacy, but there is lack of a common interface to use the available data o build new systems.
Motivation
People are creating and maintaining great data sources of social information. These data are available in different systems. However, there’s no way to build new systems to use this data split among all these sites. E.g., you moved to a new city and are looking for friends living close to you. All this data should be available to the user, so it support the development of new application, like personal agents to help users in specific tasks.
Goal
Build a common sense platform to extract data from social network web applications. The data will be available to support the development of new application, like personal agents to help users in specific tasks.
Methodology
Create a multi-agent system platform. Each social network application has an agent that knows how to extract information from the web application database. There’s an aggregated database that stores minimal information about users contacts (e.g. ids in the different systems.)
New applications can use the social network agents to get information, or just use a proxy agent that access all the other agents looking for the requested data (e.g. address, birthday, skills, work).
In the future the agents can evolve to support actions. They will be sensor and actuators working in the user behalf. The users can use them to create more sophisticated applications, so transforming the social networks from toy web apps into work web apps.
A cool example of how this application can be used is a birthday remainder. This is the subject of the b-day agent application. I’ll talk more about it soon.

