Digital Phenotyping: Making Sense of Longitudinal Health Data

This tutorial, "Digital Phenotyping: Making Sense of Longitudinal Health Data", will share the skills necessary to succeed in digital phenotyping research, covering the end-to-end research pipeline from data collection and processing to advanced analysis and scaling for real-world applications. Attendees will delve into topics such as phenotyping multidimensional and longitudinal behaviors from wearable sensors, incorporating ecological momentary assessments (EMAs) into research protocols, and analyzing smartphone text messaging data. Designed for the intended audience of researchers, data scientists, and practitioners in computer science, biomedical informatics, public health, and related fields, the tutorial is suitable for participants with varying levels of experience, offering a mix of foundational knowledge and hands-on Python-based exercises. By participating, attendees will acquire new knowledge and perspective, including the ability to conduct comprehensive digital phenotyping research. This includes extracting and analyzing complex longitudinal health data, developing robust study designs, addressing key challenges such as data completeness, and applying findings to critical areas like mental health, chronic disease management, and remote patient monitoring. By the end of this tutorial, attendees will be equipped to implement and scale digital phenotyping methods in their own research and applications.