The Challenge
The research group needed a single mobile interface that could visualise real-time sensor telemetry, trigger smart-home automations, and log participant interactions for behavioural research — across iOS and Android, without requiring technical knowledge from participants.
Approach and Execution
The app subscribes to an MQTT broker for real-time device events (sub-100ms latency). I built a declarative UI configuration system so researchers could redesign the control interface through a JSON schema without touching native code. All interactions were time-stamped and streamed to InfluxDB for analysis.
Results
Key Insight
The declarative UI pattern was the right call — researchers iterated the study design 11 times during the 8-month period. Without it, each iteration would have required a new app build and review cycle. The config-driven approach reduced iteration time from weeks to hours.
Outcome
The system was deployed across research households in three countries and generated the telemetry dataset used in two published academic papers on occupancy-aware automation.
React Native · MQTT · WebSockets · InfluxDB · Grafana