Elderly Monitoring Platform

A privacy-focused, self-hosted elderly monitoring platform that combines sensors, camera events and backend services to detect activity gaps and support family safety workflows.

This project is designed as a practical startup-style prototype: a local-first monitoring system that uses event-driven architecture, real-time alerts and a private server deployment model instead of exposing sensitive household data to public cloud services by default.

Why this project exists

Many elderly monitoring systems are either too simple, too expensive or too dependent on cloud platforms. This project explores a self-hosted alternative that can monitor activity patterns inside a home using inexpensive sensors and local infrastructure.

The goal is not to watch someone constantly. The goal is to detect meaningful risk signals such as no motion for too long, door activity at unusual times, medication cabinet inactivity or camera person-detection events that may require attention.

Core concept

The platform collects events from home devices, normalizes them through backend services and presents them through dashboards and alert workflows.

Example event flow:

Zigbee Motion Sensor / Door Sensor
        |
        v
Zigbee2MQTT + Mosquitto
        |
        v
Hardware Bridge Service
        |
        v
FastAPI Backend
        |
        v
PostgreSQL + Redis
        |
        v
Dashboard / Alert Logic / Family Review

Camera event flow:

Tapo RTSP Camera
        |
        v
Frigate Detection
        |
        v
MQTT Event
        |
        v
Frigate Bridge
        |
        v
FastAPI Backend
        |
        v
Dashboard / Event Timeline

Current prototype stack

• FastAPI backend
• PostgreSQL database
• Redis for event/queue style workflows
• Docker Compose deployment
• Zigbee2MQTT for Zigbee device integration
• Mosquitto MQTT broker
• Tapo indoor camera with RTSP
• Frigate for local camera person detection
• Sensor/event bridge services
• Local dashboard and API-first architecture

Pilot hardware

• Sonoff Dongle Lite MG21 Zigbee/Thread USB dongle
• Mini smart door sensors
• Wireless motion sensor
• Tapo indoor camera
• Planned wearable integration through BLE bridge concepts

Example monitored events

• door_opened
• door_closed
• motion_detected
• motion_cleared
• camera_person_detected
• medication cabinet activity events
• device online/offline health events

Key product features

• Event-driven backend design
• Local-first architecture
• Real-time activity event ingestion
• Sensor and camera event normalization
• Dashboard-ready API structure
• Docker-based deployment
• Privacy-conscious design
• Future-ready architecture for mobile alerts, wearable data and AI summaries

Repository purpose

This public repository is a sanitized portfolio version of the project. It focuses on architecture, documentation, sample event formats and product thinking without exposing private household data, internal server configuration or sensitive credentials.

Portfolio value

This project demonstrates:

• Backend system design
• FastAPI API architecture
• PostgreSQL-backed event modeling
• Docker-based service orchestration
• IoT and MQTT integration
• Camera event pipeline design
• Real-world product thinking
• Privacy-focused self-hosted deployment planning

Status

Active prototype / startup-style project.

The private working implementation runs in a WSL/Linux development environment and is designed for later deployment on a self-hosted server.

Roadmap

• Add public sanitized API examples
• Add dashboard screenshots with fake data
• Add mobile notification workflow design
• Add wearable event bridge design
• Add inactivity rule engine examples
• Add family notification escalation logic
• Add deployment guide for homelab server environments
• Add redacted sample reports and event timelines

Security and privacy

This project deals with sensitive household safety data. Public examples must use fake names, fake device IDs and sanitized event payloads.

Do not publish:

• Real home addresses
• Real family names
• Camera screenshots from real homes
• Internal IP addresses
• Real device serial numbers
• API keys
• .env files
• private server credentials