Analysis of Chronic Diseases and Air Quality

This repository contains code for our Data 102 Final Project.

Group: Janise Liang, Eduardo Madrigal, Sheryl Muttath, and Nandita Radhakrishnan

Introduction

We evaluated two research questions, which has been split up into separate analysis notebooks:

1. analysis/question1.ipynb: Correlation between diabetes prevalence and various health-affecting factors such as exercise, obesity, and age. Used methods for multiple hypothesis testing, including correcting for False Discovery Rates (FDR) with Bonferroni correction, and Family-Wise Error Rates (FWER) with Benjamini-Yekutieli.
2. analysis/question2.ipynb: Causal effect (if any) of PM 2.5 concentration on COPD (chronic obstructive pulmonary disease) prevalence. Used methods for causal inference using inverse-propensity weighting (IPW) on smoking, a confounding variable.

Installation

Install all required packages to run the notebooks using pip install -r requirements.txt.

• This includes an API package called sodapy for downloading CDC data.