Cassio Melo

As I reached the half period of the project duration, I’m posting an overview of what was done so far. Its convenient to give a brief introduction about Social Network Analysis (SNA). Graphs are intuitive representations of  networks. Basically, it comprises of nodes (or vertices) conected by links (or edges) representing a presence and absence [sometimes also intensity] of a relationship between nodes. Understanding those relationships may provide many powerfull insights about patterns of interaction of things, like unveiling hidden patterns of diseases outbreaks, discovering best routes, patterns of interaction of particles in Physics, protein suceptability for combination, componentization of Java classes, positional roles of people in a company, among many, many others. Almost everything can be modeled as a graph and this is why they are so amazing. Most of the study in Social Network Analysis focuses on the problem of communities (or cluster) identification — group of nodes with more similar characteristics among them than with the others. In a structural sense, they are nodes which are highly connected inside the community and less conected with nodes outside de community.
Social Network Analysis was not really new to me, but I’ve never went further in their theoric fundations. It’s very interesting although I’d prefer focusing on its applications/implications to business and society. In the current project at INRIA I’ve crawled data from Twitter in order to create a sample for the community detection (or clustering) algorithms. Here it comes the first version of an software for community detection in Social Networks.Netcluster (yes, I’m aware of the sourceforge project with the same name), has some SNA specific functionalities like graph clustering algorithms, metrics and crawlers for social networks. It allows you for example, to see how different algorithms behaves when applied to a network, compare and analyse the results. For visualization, I used the Java Universal Graph Framewok, a very promissing api for working with graphs in Java.  

  

I’m going to release it as an OS project as soon as I can and post more about its improvements.  

Salut!

Just a note to say that I moved to Paris this week. This is part of an intership program of INRIA with my university, UFPE. I’m working in a social networks analysis project with Marie-Aude Aufare and Yves Lechevallier. I’m still adaptating to the cold (from 28 to 5 degrees Celsius) and to the French way of life. I’ll let you know about the project progress here, stay tuned!

I’m glad to announce my new personal page, much cleaner and nicer than the former. I’m sorry for not giving a Portuguese version of this site. Although Brazilians are also my intended audience, I was having serious problems to synchronize the content for both languages (actually, I’m lazy). Moreover, I believe that most Brazilians can read in English while the inverse is more complicated.

I don’t have the intention to make it a large-accessed site. Instead, it will serve as a repository to my projects, articles and publications. It will help to communicate my ideas and I hope to meet interesting people through this homepage. Be my guest!