SAS Correlation Heatmap for Cancer Alley Thesis Analysis

This repository contains SAS code to generate a professional, annotated correlation heatmap from environmental and health data analyzed in my M.S. thesis at Tulane University School of Public Health and Tropical Medicine.

The thesis, titled "Cancer Alley: Industrial Emissions and Cancer Incidence in Louisiana," examines census tract-level associations between toxic releases from industrial facilities (using EPA Toxic Release Inventory data) and cancer incidence (Louisiana Tumor Registry), while adjusting for socioeconomic and behavioral risk factors.

Project Overview

This heatmap visualizes pairwise Pearson correlations among key variables, including:

• Industrial exposure proxies (CDens, RDens, NearF_D, NearR_D, SF_D)
• Socioeconomic factors (GINI, Poverty, MedIncome)
• Demographic factors (Black)
• Behavioral risk factors (Smoking, Binge, Obesity, CheckUps)

The visualization is designed for clarity in reports and publications, with centered numeric labels, a clean color gradient, and minimal axis clutter.

Files

• correlation_heatmap.sas – Complete, reproducible SAS code
  • Imports CSV data
  • Computes correlations using PROC CORR
  • Reshapes output for plotting
  • Generates annotated heatmap with SGPLOT (HEATMAPParm statement)
  • Exports to PDF
• CORRELATION_HEATMAP.pdf – Example output from the code (generated December 2025)

Requirements

• SAS 9.4 or later
• Input CSV file with the variables listed in the macro variable corr_vars

Note: The actual thesis dataset contains sensitive health information and is not included. An example dataset is included.

Code Highlights

• Use of macro variable to define variable list once (avoids repetition and errors)
• Efficient reshaping of PROC CORR output using arrays
• Modern SGPLOT features: HEATMAPParm, overlaid TEXT for value labels, custom color model
• Journal-style PDF output suitable for theses and manuscripts

Output Example

Contact

Wendy Bogil
New Orleans, LA
[email protected]
LinkedIn