Identifying land use and meteorological factors associated with the ratio of personal PM2.5 exposures versus ambient concentration in a panel of college students Open Access

Abstract

Exposure to air pollution, such as fine particulate matter PM2.5, have been associated with many adverse health outcomes. A large portion of the urban population live within a close distance to major roadways. Given that traffic is a major source of PM2.5, there is an increased interest in understanding personal exposure to PM2.5 at fine spatiotemporal resolution. This study is motivated by a dataset of personal PM2.5 measurements from wearable sensors at 15 second intervals with geo-location information, and hourly meteorological data from stationary monitors every hour. We built linear regression models and random forest models for predicting the ratio between personal exposure to PM2.5 and background ambient PM2.5 levels. Both approaches identified several land use and meteorological factors, including distance to highway, traffic counts, temperature, and relative humidity. 

Table of Contents

1 Introduction……………………………………………………….…………..…………1

2 Methods…………………………………………………………………………….……3

2. 1 The Dorm Room Inhalation to Vehicular Exposure (DRIVE) Study……...…3

                       2.1.1 Personal Exposure Data………………………….…...…………..…4

                       2.1.2 Location Data………………………………………..…..………..…5

                       2.1.3 Meteorological and Background PM2.5 Data……………..…………5

2.1.3 Traffic Count Data…….....………………….……..………………..6

2.2 Development of the Analytic Dataset…….....………………...….…………...6

2.3 Derived Variables…….....…………….………………………….…………...8

2.4 Data Analysis...…….....……………………………………….…………......10

2.5 Prediction Grid Calculations...…….....………………………………………11

3 Results...…….....……………………………………… ………………………………12

3.1 LR Model...…….....…………………………………….……………………13

3.2 RF Model...…….....………………………………….………………………14

3.3 Predictions based on LR and RF Models………….…………………………15

4. Discussion……………………………………………………………………………..18

4.1 Limitations………………………………….………………………………..18

4.2 Future Work……………………………….………..………………………..19

About this Master's Thesis

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Rollins School of Public Health
Department	Biostatistics
Subfield / Discipline	Biostatistics - MPH & MSPH
Degree	M.P.H.
Submission	Master's Thesis
Language	English
Research Field	Statistics Biology, Biostatistics Environmental Studies
Keyword	Spatial Statistics Machine Learning Environmental Health Random Forest Air Pollution Linear Modelling Predictions
Committee Chair / Thesis Advisor	Howard Chang, Emory University
Committee Members	Donghai Liang, Emory University

Last modified

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	Identifying land use and meteorological factors associated with the ratio of personal PM2.5 exposures versus ambient concentration in a panel of college students ()	2020-04-15 14:27:24 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions