In this project, we propose a probabilistic modeling approach to represent the speed- density relationship of pedestrian traffic. The approach is data-driven and is motivated by the presence of high scatter in the raw data I have been provided with (tracking infor- mation about all pedestrians in a train station corridor during working-days). By relaxing the main assumptions of the deterministic speed-density relationship commonly known in traffic engineering, we show the relevance of accounting for both dynamics and hetero- geneity in the speed-density fundamentals. This work is an introduction to data-driven mixture modeling in pedestrian field that could be developed in future works. It presents different tracks which produce interesting results about pedestrian traffic models, and a new vision of latent-class modeling in speed-density and pedestrian behaviors in a train station distribution corridor.