I'm interested in the fields of data mining and machine learning in general and in unsupervised learning and cluster analysis in particular.

In the past, I worked on several aspects related to the management and analysis of moving object trajectories. I studied the problem of efficient on-the-fly sampling of continuous trajectory data streams during my research internship at Télécom ParisTech and for my Ph.D. thesis, that I conducted in the same institute, I worked on clustering network-constrained trajectories, mainly using graph partitioning techniques (e.g. modularity-based community detection) and co-clustering.

Currently, I continue to work on mining mobility data although in a different context. As a postdoctoral researcher at Ifsttar, I'm involved in the Mobilletic project in which we're working on mining smart card logs (bus and subway trips) in order to analyze travel behavior. For instance, we successfully used generative approaches involving mixture models (mixture of unigrams, mixture of Poisson distributions, etc.) in order to reveal frequent passenger temporal patterns and characterize station activities. At present, we're trying to study multimodality aspects (i.e. transfer behavior) in public transit using data visualization tools.

I would love to work on other topics such as mining social network data (if only Square-Enix would accept to provide me with some FFXIV data), Web mining, and healthcare data (the "grandma has Alzheimer effect"). If you're working on one of those topics and on the lookup for someone curious and motivated, don't hesitate to get in touch!

List of Publications