Skip to content
← Back to work

AirSense

An IoT web app that visualises live air quality in rooms at NTNU Gjøvik — temperature, humidity and CO2 collected from micro:bit sensors.

AirSense indoor air quality dashboard shown on a MacBook mockup

01

Overview

Project type
IoT & web application
Year
2025
Role
Frontend & web app development (shared lead with Saif Ali Rana), contributions to the physical model design
Team
4 — Emma Martine Troseth Glein, Mina Cecilie Mogen Øien, Saif Ali Rana and me
Technologies
micro:bit, Node-RED, MQTT, JavaScript, HTML/CSS
Tools
Figma, GitHub

GoalBring attention to indoor air quality by collecting live environmental data and making it readable at a glance in a web app.

02

The challenge

Poor air quality in classrooms and study rooms at NTNU Gjøvik is invisible — students and staff have no way of knowing when CO2 levels, humidity or temperature are affecting concentration and comfort.

Target usersStudents and staff using rooms at NTNU Gjøvik.

03

Research & understanding

  • Looked into recommended thresholds for CO2, humidity and temperature in learning environments to set meaningful states in the interface.
  • Tested sensor placement and reading intervals to find a balance between responsiveness and noise.

04

Process

  1. 01

    Concept

    Framed the problem around invisible air quality in shared rooms on campus.

  2. 02

    Hardware

    Wired temperature, humidity and CO2 sensors and processed readings locally on a micro:bit.

  3. 03

    Data pipeline

    Sent readings wirelessly over WiFi into a Node-RED based platform.

  4. 04

    Web app

    Built the interface that presents live values and their state, with Saif.

  5. 05

    Physical model

    Contributed to the design of the physical model used in the exhibition.

  6. 06

    Presentation

    Demoed the running system and gathered feedback.

05

Design

  • Prioritised a single glanceable status per room over dense charts — the app answers 'is the air okay right now?' first.
  • Used colour states tied to real thresholds so the data is interpretable without prior knowledge.

06

Development

  • Sensor data is processed on the micro:bit and pushed wirelessly to a Node-RED flow that normalises and forwards it.
  • The web app subscribes to the live stream and renders current values with history for context.
  • micro:bit
  • Node-RED
  • MQTT
  • JavaScript
  • HTML/CSS

07

Final solution

A working end-to-end IoT system: physical sensor unit, Node-RED data platform and a web app where anyone can follow temperature, humidity and CO2 for a room in real time.

The project received good feedback and gave the team hands-on experience with a full hardware-to-web pipeline.

08

Reflection

What worked
Splitting hardware and web app work while keeping a shared data contract kept the team moving in parallel.
What did not work
Early sensor readings were noisy and needed smoothing before they were usable in the interface.
What I learned
How to work in a team across hardware, data and frontend, and how much interface clarity depends on trustworthy data.
What I would improve
Add historical trends per room and alerts when values stay above threshold for longer periods.
My contribution
Shared main responsibility for the web app and helped shape the design of the physical model.

Next project

AI literacy test

AI Literacy Quiz result screen showing a score of 8/10, advanced level, and topic breakdown