About This Project

Methodology and Technical Details

Author

Aaron Chafetz | Digital Service at CMS

Project Purpose

This website serves as a practical demonstration of Quarto’s capabilities for data analysis and reporting, specifically tailored for analysts working with healthcare and government data.

By comparing two versions of the same Medicare enrollment analysis, you can see firsthand how Quarto’s features enhance reproducibility, automation, and presentation quality.

Why This Matters for CMS Analytics

Working with Medicare and Medicaid data requires:

  • Reproducibility: Auditable analysis processes
  • Automation: Regular reporting cycles (monthly, quarterly, annual)
  • Transparency: Stakeholders need to understand methodology
  • Flexibility: Multiple output formats for different audiences

Quarto addresses all these needs in a single, open-source framework.

Data Source

All analyses use public Medicare monthly enrollment data from Data.CMS.gov, the official open data portal for the Centers for Medicare & Medicaid Services.

This real-world dataset demonstrates how Quarto handles:

  • Large government datasets
  • Time-series analysis
  • Geographic variations
  • Reproducible reporting requirements common in healthcare analytics

Technical Implementation

Quarto Components Used

1. YAML Configuration

The YAML header controls document behavior and appearance:

  • Dynamic date: Automatically updates to the render date
  • Dynamic subtitle: Can pull from parameters or calculations
  • Code rendering defaults: Control whether code is shown or hidden by default
  • Output formats: HTML, PDF, Word, presentations, and more

2. Code Chunks

Analyses are written in R (though Quarto supports Python, Julia, and Observable):

  • Execute data processing and visualization
  • Generate tables and statistics
  • Create interactive elements

3. Markdown and Inline Code

Narrative text integrates seamlessly with code results:

  • Explain methodology and findings
  • Include inline calculations (e.g., “The enrollment increased by r growth_rate%”)
  • Format professional reports

Advanced Features Demonstrated

Code Folding

In the fancy demo, code is hidden by default but accessible via a “Show code” link in the upper right corner. This keeps the document clean for non-technical readers while maintaining full transparency.

Implementation:

yaml
format:
  html:
    code-fold: true

Table of Contents

Automatic navigation based on document headers improves usability for longer reports.

Implementation:

format:
  html:
    toc: true

Custom CSS Styling

External CSS files allow consistent branding across multiple reports without cluttering your Quarto document.

Implementation:

format:
  html:
    css: styles.css

Parameterized Reports

The params feature in YAML enables template-based reporting:

Implementation:

params:
  state: "California"
  year: 2024

These parameters can be referenced throughout the document and changed programmatically to generate reports for all states, years, or other dimensions automatically.

Example use case: Generate 50 state-specific reports with a single command:

# R code to render for all states
states <- state.name
for (state in states) {
  quarto::quarto_render(
    "analysis.qmd",
    execute_params = list(state = state),
    output_file = paste0("report_", state, ".html")
  )
}

Getting Started with Quarto

Installation

  1. Install Quarto from quarto.org
  2. Use with VS Code, Jupyter, RStudio, Positron, or any text editor
  3. Render documents with quarto render or IDE buttons

Resources

Repository

This project is hosted on GitHub at https://github.com/achafetz/dsac_quarto_demo/. Feel free to explore the code, fork the repository, or contribute improvements.

All data are public and sourced from Data.CMS.gov

Disclaimer: The findings, interpretation, and conclusions expressed herein are those of the authors and do not necessarily reflect the views of Centers for Medicare and Medicaid Services. All errors remain our own.