Faculty in the Mobile and Wearable Data Science Hub develop statistical and machine learning methods for extracting insights from continuous, high-frequency, person-generated data from mobile apps, wearable sensors, and digital platforms. Work focuses on real-time health monitoring, precision/personalized cancer risk modeling, and personalized interventions, with methods spanning signal processing, functional data analysis, dynamic treatment regimes, and privacy-aware learning.
Hub Faculty
Chongzhi Di
Accelerometry & Wearable Sensor Analytics · Functional Data Analysis · mHealth Interventions
Chongzhi Di develops statistical methods for mobile and wearable health data, focusing on functional and longitudinal modeling for high-frequency accelerometer signals and other sensor streams. Motivated by ancillary studies of the Women’s Health Initiative, his group has advanced analysis of device-measured physical activity, including a new metric for summarizing raw tri-axial accelerometry data and functional/compositional tools for relating activity patterns to health outcomes. He also collaborates on mobile-health studies using smartphone-based behavior interventions and wearable ECG monitoring to improve measurement, monitoring, and evaluation in real-world settings.
Jeff Leek
Digital Phenotyping · Inference After Prediction · Reproducible Analytics
Jeff Leek leads Fred Hutch’s integrated data science enterprise as vice president and chief data officer, working to ensure researchers can access field-leading computational resources, shared data services, and software tools and reproducible workflows for working with biomedical data. As a biostatistician, he develops statistical methods, software, and large-scale data resources that support reproducible analysis and valid inference in modern data settings, including studies that rely on algorithmically derived measures and predictions. In the mobile and wearable context, this work helps teams move from raw digital signals and linked clinical data to reliable, scalable analyses that can be reused across studies and translated into usable tools and services.
Yingqi Zhao
Dynamic Treatment Regimes · Precision Medicine · Disease Surveillance Using Real-World Data
Yingqi Zhao develops statistical and machine-learning methods for personalized medicine using longitudinal and real-world health data, including electronic health records. Her work on dynamic treatment regimes and precision treatment/prevention supports adaptive strategies that update as a patient’s information evolves. She also develops methods for disease screening, public health surveillance, and clinical trial design, with applications spanning early cancer detection, cancer immunotherapy trials, and chronic disease settings such as Type 2 diabetes and childhood obesity surveillance.