This paper reports four cases of action sensing and data analysis concerning action sports, i.e. skateboard, inline skate, and bicycle motocross (BMX). All cases employs inertial sensors with GPS to synchronize time series data collected from the independent multiple sensors on parts of athlete's body. All the cases also define similarity to classify running data and to identify important factors for successful maneuver. The first case is related to identification of types of sports and skill levels on halfpipe. The second case is also related to identification problem on big air ramp. The third case analyzed aerial maneuvers of skateboard on halfpipe in detail. The forth case identified important factors of incremental training to make successful aerial maneuvers.