Cycling Activity Dataset Creation and Application for Feedback Giving

Abstract

Our research aims at supporting personal cycling trainer applications in training planning and feedback giving to nonprofessional outdoor cyclists, based on a general reference. In this paper we present the created dataset. According to our present knowledge, this data collection is the first public dataset containing cycling activities recorded outdoor. Its usability for training planning and feedback giving is demonstrated through an example. The dataset is clustered according to age groups, considering distance and average speed as the two most influential features when predicting the time required for training. These clusters are then applied as references in feedback giving and goal setting.