The Problem

Every iPhone silently counts steps — but no consumer app explains the signal processing behind it, or goes further to measure gait quality: how hard you land, and whether your left and right legs are keeping the same rhythm.

What I Built

• Streams raw CoreMotion accelerometer data at 50Hz, collapses x/y/z into an acceleration magnitude signal, and detects footsteps via peak detection — the same fundamental technique used inside Apple's pedometer — computing step cadence in real time from a rolling 3-second buffer
• Computes two gait metrics from the step stream: impact (average peak acceleration per footfall, mapped to Low/Moderate/High) and symmetry (variance in alternating left/right time gaps, exposing asymmetric gait or a limp as a measurable number)
• Blends both into a live 0–100 Walk Quality Score with clean SwiftUI real-time display; fully on-device MVVM architecture with no cloud dependencies or third-party ML libraries

Impact

A from-scratch exploration of the signal processing that underpins health wearables — demonstrating that the "magic" of step counting is peak detection on an accelerometer signal, and extending it into clinically meaningful gait quality metrics.