This paper aims to investigate rider braking behaviors using a dataset of braking maneuvers derived from naturalistic riding data. Each braking event was fully characterized with experimental data. A set of descriptive parameters was defined to capture relevant information of the braking event and to facilitate the clustering process of braking behaviors. Naturalistic data of 5 riders were automatically processed to identify and characterize the braking events based on the given set of parameters. A preliminary descriptive analysis was performed to verify the presence of macro behaviors of riders. Subsequently, a Principal Component Analysis was performed to reduce problem dimensionality and support the cluster analysis on the dataset of a rider. The results indicated that a macro classification of riders is possible also based on a descriptive analysis. Nonetheless a cluster analysis sharply identified different behaviors of the rider, and thus provided a more solid basis for comparison of behavior among riders. In addition, the clusters revealed quantitative data that will be useful for the development of assistive systems.