Survival Analysis Practice

This repository presents a complete survival analysis of HIV patients who experienced failure with zidovudine therapy, using the aids and aids.id datasets from the JM package. The project showcases not only advanced modeling but also careful data exploration, assumption checking, and clear interpretation using reproducible R workflows.

---

Repository Structure

survival_analysis_practice/
├── LICENSE
├── README.md
├── rendered_docs/                  # Contains all the processed reports
│   ├── Appendix_Rajaram.pdf
│   ├── Codes_Rajaram_1.pdf
│   ├── Codes_Rajaram.pdf
│   └── Project_Rajaram.pdf
├── RMD/                            # All the RMarkdown files for code, analysis, and interpretation
│   ├── Appendix_Rajaram.rmd
│   ├── Codes_Rajaram_1.rmd
│   ├── Codes_Rajaram.rmd
│   └── Project_Rajaram.rmd
└── survival_analysis_practice.Rproj  # RStudio project file

---

Document Descriptions

Project_Rajaram.rmd

Main interpretive report

• Explores patient-level and clinical variables
• Fits Kaplan-Meier curves, Cox models (time-independent and time-dependent), and AFT models (Weibull, Log-normal)
• Includes PH assumption checks, functional form checks, and influence diagnostics
• Tables and figures are generated using external scripts (from Codes_Rajaram.rmd)
• Presents tidy summaries, interpretation of outputs, and conclusions

Codes_Rajaram.rmd

Code script for core analysis

• All code for:
  • Data preprocessing
  • Model fitting (KM, Cox, AFT)
  • Assumption diagnostics
  • Summary tables and visualizations
• Modularized and structured to support the main report (Project_Rajaram.rmd) and appendix

Appendix_Rajaram.rmd

Supplementary outputs

• Additional figures and tables referenced in the main report
• Includes diagnostics and stratified plots to support interpretation
• All outputs generated from Codes_Rajaram.rmd

Codes_Rajaram_1.rmd

Standalone full code notebook

• Contains all code chunks used in Codes_Rajaram, Project_Rajaram and Appendix_Rajaram in a single file
• Suitable for full replication of the entire workflow
• Does not include narrative interpretations

---

Skills Demonstrated

• Survival analysis: KM curves, Cox PH models, AFT models
• Time-varying covariates and PH assumption testing
• Residual diagnostics and model validation
• Tidyverse-based data processing and broom-style model outputs
• Clean modular RMarkdown structure for reproducibility
• Tidy tables (gt), visualizations (ggplot2), and interpretive scientific writing

---

Rendered Reports

Available in the rendered_docs/ folder:

• Project_Rajaram.pdf: Main interpretive analysis
• Codes_Rajaram.pdf: Modular code and generation of outputs
• Appendix_Rajaram.pdf: Additional plots, tables, and figures
• Codes_Rajaram_1.pdf: Self-contained code-only notebook

---

How to Use This Repository

1. Clone this repo and open survival.Rproj in RStudio
2. Navigate to the RMD/ folder
3. Knit Project_Rajaram.rmd to produce the main report
4. Use Codes_Rajaram.rmd to generate and update output tables/figures
5. Use Codes_Rajaram_1.rmd as a full standalone code reference