Indoor Climate System

Overview

A full-stack embedded monitoring system that captures real-time indoor air quality data, transmits it over MQTT, stores it in the cloud, and displays it on a locally hosted dashboard.

Motivation

Poor indoor air quality often goes unnoticed. High CO₂ levels can reduce concentration, elevated particulate matter affects respiratory health, and improper humidity encourages mould growth. This project was built to make these invisible factors visible — providing continuous, data-driven insight into the indoor environment.

Sensor Box

The core of the system is a custom sensor box that measures:

CO₂ (ppm) — carbon dioxide concentration
Temperature (°C)
Particulate Matter (PM2.5 & PM10, µg/m³) — fine and coarse airborne particles
Humidity (RH%) — relative humidity

Sensor readings are sampled at regular intervals and packaged into structured messages for transmission.

Architecture

Sensors → Raspberry Pi → MQTT Broker → AWS IoT Core → DynamoDB
                ↓
         Local Web Dashboard

Data acquisition — Sensors connected to a Raspberry Pi collect environmental readings via C++ drivers.
Message transport — Readings are published over MQTT, providing lightweight and reliable delivery.
Cloud storage — AWS IoT Core ingests the messages and routes them into a DynamoDB table for persistent, queryable storage.
Visualization — A web dashboard hosted on the Raspberry Pi fetches historical and live data, rendering it in time-series charts.

Tech Stack

Layer	Technology
Hardware	Raspberry Pi, SCD30 (CO₂), SPS30 (PM), DHT22 (temp/humidity)
Firmware	C++ with Linux system APIs
Transport	MQTT (Mosquitto broker)
Cloud	AWS IoT Core, DynamoDB
Dashboard	HTML/JS served from the Pi

Key Learnings

Designing reliable sensor polling loops in C++ on embedded Linux
Structuring MQTT topic hierarchies for scalable IoT data
Configuring AWS IoT rules to route messages into DynamoDB
Building a lightweight dashboard that works on a low-power device