The Berkeley Telemonitoring Project is a framework for building Android-based applications for remote data collection, analysis, and feedback. We have expanded this framework to support needed functionality for a long-distance running coaching application, as injury in the lower extremities is prevalent in runners. Cadence (steps per minute) has been identified as a key metric to reduce injury and improve running performance. Thus, the primary areas of work for this expansion include estimators for cadence and speed from accelerometer data based on scientifically validated algorithms, along with additional modifications to the framework to include data analytics techniques and improved Bluetooth connection handling. Finally, we developed a sample coaching application that includes these framework components to be tested by experienced long-distance runners in an approved pilot study. As this framework will soon become available to developers and medical professionals in the open-source community, we hope that our extensions to the codebase will benefit others who endeavor to create telemonitoring applications.