Learning digital signal processing using an interactive Jupyter notebook and smartphone accelerometer data

Digital signal processing is a valuable practical skill for the contemporary physicist, yet in physics curricula its central concepts are often introduced either in method courses in a highly abstract and mathematics-oriented manner or in lab work with little explicit attention. In this paper, we present an experimental task in which we focus on a practical implementation of the discrete Fourier transform (DFT) in an everyday context of vibration analysis using data collected by a smartphone accelerometer. Students are accompanied in the experiment by a Jupyter notebook companion, which serves as an interactive instruction sheet and a tool for data analysis. The task is suitable for beyond-first-year university physics students with some prior experience in uncertainty analysis, data representation, and data analysis. Based on our observations the experiment is very engaging. Students have consistently reported interest in the experiment and they have found it a good demonstration of the DFT method