wearablecompute is an open source Python package containing over 50 data and domain-driven features that can be computed from wearables and mHealth sensor data.

Data Science Project - Predicting glucose levels with data collected by non-invasive wearable device

Parkinson's detection from 3 seconds of voice — wav2vec2-XLS-R embeddings + LR, CV AUC 0.972, cross-corpus validated across Italian and English datasets

A highly customizable web dashboard for analyzing and exporting multi-metric Empatica Embrace biometric data and activity classifications.

Passive neurological screening via TinyML — voice tremor, gait, typing & touch biomarkers

Time-frequency analysis and a structure-preserving colour encoding for Parkinson's disease detection from PaHaW handwriting — evaluated under repeated nested cross-validation with the Nadeau–Bengio corrected t-test.

Free-of-cost web platform for longitudinal cognitive monitoring and clinician-guided rehabilitation in multiple sclerosis. Features 35 adaptive cognitive tasks across 6 domains, digital biomarker extraction, bilingual Bengali/English support, and coordinated patient-clinician workflows.

Reproducible analysis code and aggregated outputs for a VR-based cognitive screening tool with integrated eye tracking (AD, MCI, and controls).

Benchmarking open-source PPG foundation models (Pulse-PPG, PaPaGei-S) for biological age prediction on clinical data

Digital neuroscience platform for quantifying brain stability, deviation, and resilience using longitudinal baseline modeling.

MATLAB code for: "Photoplethysmography Acceleration Indices as Digital Biomarkers of Cardiometabolic Health"

Native Android app for people with multiple sclerosis. Weekly self administered five test battery (gait, dexterity, vision, cognition, speech), tracked on device. Kotlin + Jetpack Compose + Room. No internet permission, no cloud.