Wearable sensors have made a huge impact on a wide range of industries since their inception. Due to their variety of form factors, these sensors are becoming rather popular for daily casual commercial use as well. One of the many widely used functionalities of wearable sensors is human activity recognition. These sensors are capable of using data to ascertain the activity being performed by a user, which they in turn use to provide useful insights. However, there are several challenges involved in designing these sensors. One of the factors that impacts the sensor’s efficacy in recognizing activity is the position on the human body at which the sensor is mounted. Although, often assumed to be a benign influence, mounting positions may severely effect the quality of data that is utilized by the activity recognition classifier thereby affecting prediction accuracy. The goal of this project is to conduct an empirical study wherein data is collected, processed and and analyzed for its impact on activity recognition using machine learning techniques.