2 Quick Start

This chapter walks through installation, environment setup, and running your first analysis end-to-end.

2.1 Prerequisites

Requirement	Version	Notes
R	>= 4.2	With `renv` for package management
Python	>= 3.12	Managed via `uv`
Quarto	Latest	CLI for manuscript rendering
gfortran	Any	Required by `frailtypack` (`brew install gcc` on macOS)
uv	Latest	Python package manager

2.2 Installation

# Clone the repository
git clone https://github.com/htlin222/survival-pipe.git
cd survival-pipe

# Bootstrap R and Python environments
bash setup.sh

# Verify everything is ready
bash verify_environment.sh

First-Time Setup

setup.sh installs R packages via renv::restore() and Python dependencies into a .venv managed by uv. Run it once after cloning.

2.3 First Analysis Walkthrough

2.3.1 1. Initialise a project

make init PROJECT=my-study

This creates the standard directory tree under projects/my-study/.

2.3.2 2. Simulate data (for demo purposes)

make simulate SCENARIO=1 PROJECT=my-study

Scenarios 1–7 generate datasets with different survival complexities (standard, competing risks, recurrent events, etc.).

2.3.3 3. Run diagnostics

make diagnose PROJECT=my-study

Validates data structure, runs all six diagnostic checks, and produces a routing report.

2.3.4 4. Auto-select and run analysis

make analyze-auto PROJECT=my-study

The diagnostic router reads the report and executes the appropriate analysis module.

2.3.5 5. Generate figures and manuscript

make figures PROJECT=my-study
make manuscript PROJECT=my-study
make render PROJECT=my-study

2.4 Makefile Targets at a Glance

Target	Description
`make help`	List all available targets
`make setup`	One-time environment bootstrap
`make verify`	Check environment health
`make init`	Create a new project
`make simulate`	Generate simulated data
`make diagnose`	Run all diagnostic checks
`make route`	Diagnose + recommend analysis
`make analyze-auto`	Auto-select and run analysis
`make figures`	Generate 300 DPI figures
`make manuscript`	Build Table 1 and results
`make render`	Render Quarto to PDF + DOCX
`make qa`	Run QA validation
`make all`	Full pipeline end-to-end
`make clean`	Remove generated outputs

Clean Target

make clean removes all generated outputs (diagnostics, analysis, figures, manuscript) but preserves raw data. It uses rip (not rm) for safe deletion.