Residential IAQ & Fireplace Emissions

Modifying generic IoT hardware for a multi-site deployment, encompassing ethical administration, high-frequency data analysis, science communication, and a peer-reviewed publication.
Hardware Adaptation Stakeholder Management Data Analysis Science Comm

While my primary LoRaWAN sensor network was designed for educational infrastructure, I modified and deployed my generic Indoor Air Quality (IAQ) sensing device into a residential setting to investigate the impact of amenity indoor fireplaces on household air quality.

This project required more than just hardware engineering; it demanded rigorous stakeholder management, formal research administration, and the coordination of a multi-site deployment across eight private households.

The high-frequency data collected during the five-month winter period was rigorously analysed, resulting in a first-author publication in the peer-reviewed journal Energy Research & Social Science and widespread national media coverage.

1. Hardware modification & Firmware adaptation

Deploying sensors in affluent, private households requires a different approach than deploying them in public school infrastructure. The hardware had to be unobtrusive, reliable, and aesthetically acceptable for a living room environment.

  • Hardware Adaptation: I adapted my generic Atmel-based IAQ node, utilising the Plantower PMS5003 laser-scattering particulate matter sensor. The enclosure was redesigned and tested to ensure ample airflow and ventilation while maintaining a compact footprint that could be powered continuously via a standard electrical outlet.
  • Firmware for Transient Capture: To accurately capture the rapid spikes in particulate matter (PM1.0, PM2.5, and PM10) associated with lighting a fire, I modified the embedded firmware. The device was programmed to sample and transmit data precisely every 11 minutes (over 130 times daily), ensuring no transient emission events were missed.
The modified sensing node deployed in the residential households, adapted from my generic IAQ hardware architecture.

2. Stakeholder management & Ethical administration

Conducting hardware research inside private homes introduces profound logistical and ethical complexities that go beyond standard engineering tasks.

  • Project Proposals & Ethical Clearance: I authored the formal project proposal and comprehensive technical reports required by the institutional review board. I successfully navigated the ethics committee, drafting the strict protocols needed to ensure occupant privacy, data anonymisation, and informed consent for human-environment research.
  • Multi-Site Deployment & Coordination: I managed the physical deployment across eight different households in Stellenbosch. This required coordinating installation schedules with homeowners, ensuring proper sensor placement relative to open and enclosed fireplaces, and maintaining continuous communication to manage device uptime throughout the winter season.

3. Data analysis & Environmental correlation

The devices generated a dense, high-frequency dataset that required complex pre-processing to extract meaningful insights regarding human behaviour and indoor air quality.

  • Data Processing Pipeline: I developed Python scripts to clean the dataset, removing invalid measurements and calculating highly specific 24-hour sliding window averages to benchmark the data against strict WHO and EPA exposure guidelines.
  • Ambient Weather Integration: To prove that colder weather drove fireplace usage and subsequently worsened indoor air quality, I integrated external meteorological data. By pulling hourly ambient temperature data from the NOAA Environmental Monitoring System, I programmatically correlated outdoor weather patterns with indoor PM2.5 spikes.
Data analysis demonstrating the negative linear correlation between outdoor ambient temperatures and indoor PM2.5 spikes during fireplace use.

4. Real-world behavioural impact

Data is only valuable if it drives change. During the experiment, my automated data analysis flagged dangerously high PM2.5 levels in one specific household. By presenting these data-driven findings directly to the stakeholders, the occupants immediately ceased indoor fire-making, resulting in a stark, immediate drop in particulate matter.

Time-series data proving the immediate, drastic reduction in indoor PM2.5 levels after occupants were informed of the sensor readings.

5. Science communication & Media engagement

Beyond the academic publication, the findings of this project garnered significant national media attention. Distilling dense statistical data, air quality guidelines, and complex engineering methodologies into accessible, actionable information for the general public requires high-level Science Communication.

I conducted a comprehensive press tour to break down this complex scientific information. This included live televised interviews on eNCA, and numerous radio segments on national and local stations such as RSG, Cape Talk, and various Johannesburg broadcasts. The work was also featured in several blog posts and an official Stellenbosch University press release.

Live television interview on eNCA, discussing the engineering methodology and public health findings of the residential fireplace study.


Stellenbosch University Press Release: The video below shows an extract of the soundbites recorded for the official university press release, summarising the core findings of the study for public distribution.