Skip to content

Sharing all the data pipelines and processing codes, statistical modellings, descriptive statistics, plot visualizations, and machine learning from Mahdavi & Siegel (2021) (Indoor Air) Project Miestone: 2017 - 2020 Full-length article: https://onlinelibrary.wiley.com/doi/abs/10.1111/ina.12782

Notifications You must be signed in to change notification settings

armahdavi/code-data-analytics-pipelines-statistics-ML-PM-PSD_residential-QFF

Repository files navigation

Summary

This repository summarizes all the data pipelines, data pre-processing codes, statistical models, descriptive statistics, plot visualizations, and ML modelings from Mahdavi & Siegel (2020) (Indoor Air)
Project Milestone: 2017 - 2021

PSD and Concentration: Two Important Determinants of the Health Effects of Exposure to PM

Particle size distribution (PSD) and concentration are two of the most important features of particulate matter (PM). which is responsible for a great loss of life from exposure by inhalation, dermal contact, or ingestion. Samples of PM can be found and studied in the dust collected on HVAC filters in residential buildings. Therefore, the application of HVAC filters in collecting and analyzing PM indoors is a useful and cost-effective strategy.

The Hidden Story in Our Air Filters

HVAC filters play a dual role in our homes. Not only do they purify the air we breathe by trapping harmful particles, but they also act as silent samplers, collecting a record of these airborne contaminants. This makes them a valuable resource for experts in Indoor Air Quality (IAQ), Indoor Environmental Engineering, and Health. By analyzing the dust accumulated on used filters, researchers gain insights into the particles and contaminants we may be inhaling.

This analysis technique is known as Filter Forensics. When combined with metadata from HVAC systems (airflow rate, runtime, filter efficiency), it allows for calculating long-term airborne particle concentrations. This concentration data is crucial for exposure assessments in chronic health studies. This more quantitative approach is termed Quantitative Filter Forensics (QFF).

This Repository

This repository showcases the data pipelines, pre-processing code, statistical models, descriptive statistics, and visualizations used in Phase 1 of the Mahdavi & Siegel (2020) (AS&T) paper. In this phase, filters were artificially loaded with standardized test dust samples in the lab for further extraction experiment processes. While the paper itself discusses the front-end results, this repository provides the underlying code that generates those results, allowing you to explore the analysis firsthand.

The pre-processing and logical Python codes are provided in .py (originally written in Spyder), and data visualization, statistical models, and front-end calculations are provided in .ipynb (written in Jupyter) for a guided walk-through of the data pipeline process.

About

Sharing all the data pipelines and processing codes, statistical modellings, descriptive statistics, plot visualizations, and machine learning from Mahdavi & Siegel (2021) (Indoor Air) Project Miestone: 2017 - 2020 Full-length article: https://onlinelibrary.wiley.com/doi/abs/10.1111/ina.12782

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published