In recent years, the use of bicycle has been promoted worldwide as a healthy and economical means of transportation, and bicycles are attracting attention as a means of last mile transportation in MaaS. In this research, potential comfort of the riding environment is extracted from facial expressions during the bicycle ride. By acquiring the driver’s facial expressions and scenery from a smartphone mounted on the bicycle and analyzing the emotions, the system extracts the user’s potential comfort of the riding environment and automatically annotates the locations. This can then be used for navigation and location feature mapping. In addition, by explicitly collecting user feedback on ride comfort, the system can learn latent comfort for the driving environment. In this paper, we construct a latent comfort analysis and the map generation system for locations based on facial expressions, and demonstrate it experimentally on 17 users living in the Kansai region in Japan.