Finding similar neighborhoods across cities

By Geraud Le Falher, Michael Mathioudakis, and Aris Gionis

Our friend Oliver lives in London, where he works as a consultant for a big financial company. Occasionally, he takes a business trip to another major city, to seal a major deal, and make major buck for his boss. Of course, Oliver being Oliver, he always finds the time to enjoy whatever city he happens to be. When in London, he likes to suit up and hang out in Soho, a “predominantly fashionable district of upmarket restaurants.” He would like to do that also in Rome, where he’s flying to next week, but he doesn’t know much about that city. Where is the Soho of Rome? What neighborhood of Rome is most similar to Soho? Continue reading

Overlapping community detection in labeled graphs

In a recent paper with Esther Galbrun and Nikolaj Tatti, presented in the journal of Data Mining and Knowledge Discovery, we worked on the problem of discovering overlapping communities in networks with labeled vertices. The model is motivated by social networks, where vertex labels are used to represent information about individuals, such as occupation, hobbies, preferences, etc. The hypothesis is that the vertex labels can be used to derive and explain the community structure in the network. Continue reading

From “I love you babe” to “leave me alone” — romantic relationship breakups on twitter

Our recent paper (together with Ingmar Weber and Sonia Dal Cin) studying Romantic Relationship Breakups on Twitter was accepted at the 6th International Conference on Social Informatics (SocInfo 2014). In the paper, we identified pairs of Twitter users who were in a romantic relationship and studied various psychological processes surrounding relationship dissolution. Continue reading

How graph algorithms can help to find interesting events

In our recent KDD paper, with Polina Rozenshtein, Aris Anagnostopoulos, and Nikolaj Tatti, we worked on the problem of finding events in graphs. We abstracted the event-finding problem with the following simple formulation: Given a graph with node weights and edge distances, find a subset of nodes (the event) that have large sum of weights and are well connected. In the paper we showed how to use this formulation to find interesting events in real-world datasets. Continue reading

Quantifying rappers’ rhyming skills

“Women lie, men lie, numbers don’t” – Jay-Z

Last Monday night, I posted to my personal blog a small study where I analyse the lyrics of Finnish hip-hop artists. By Tuesday night, the post had attracted over 32,000 page views, several online newspapers had reported about it, and I had given an interview to a national radio station. In this post, I will give a brief summary of what all the fuss was about. Continue reading