Findings in sports technology tell us that more and more sports monitoring equipment are being developed and extensively used for tracking fitness data in different physical conditions. In this paper, we proposed a novel IoT solution to track and utilize fitness data by employing data analysis and machine learning principles. We developed an Android Application to execute digital signal processing algorithms over real-time fitness data collected by gyroscope and accelerometer in WICED Sense Kit through Bluetooth transmission. Then after careful calculations, we translate the exercise data into quantitative data, which was analyzed by our machine learning algorithm on the Flask server deployed on AWS EC2 instance, to perform personalized recommendation. We also developed a website to serve as a social platform for users to track, explore and share their fitness data online.