There are many existing methods designed to measure physical activity in public health research. Traditionally, researchers use subjective methods like physical activity diaries and surveys to collect such data as they are relatively inexpensive. However, such methods frequently produce biased results due to the self-report nature of the data collection. In more recent years, wearable accelerometers, or actigraphy devices, have become the norm in physical activity research as they offer a noninvasive and objective measure of participants’ daily movements. Accelerometers are relatively small devices that record high frequency time series measurements corresponding to the movements of participants. Analyses of the collected data are complicated by the fact that true physical activity is contaminated by a large amount of noise. Walking is often the only form of sustained physical activity in many populations, but other common daily tasks, such as driving a car, can mimic the walking signal. In our case study, we cover some of the basic features of accelerometer-based measurements, considerations we made when designing the Indiana University Walking and Driving Study, and cover some appropriate feature extraction methods to aid in differentiating between different types of physical activity.