Rigorous, data-driven performance research by Code Evolution Lab. Each study is self-contained under studies/ with its own dependencies, benchmarks, results, and articles.

empirical-study/
  studies/
    01-n-plus-1-query/      # Prisma N+1 query benchmarks
    02-blocking-io/         # Blocking I/O prevalence + benchmarks
    03-memory-leaks/        # Memory leak detection in React/Vue/Angular
    04-loop-performance/    # Loop anti-pattern benchmarks + corpus study
    05-missing-index/       # Missing index benchmarks (PostgreSQL + Prisma)
    06-resource-leaks/      # Resource leak benchmarks + real-world corpus scan
  docs/                     # Shared writing guides, research plan
  tsconfig.base.json        # Shared TypeScript config
  package.json              # npm workspaces root

Each study folder contains:

• README.md — Study overview, methodology, quick start
• package.json — Study-specific dependencies
• src/ — Benchmark and analysis code
• results/ — Raw benchmark output (gitignored)
• content/ — Article drafts (gitignored)

# Install all workspaces
npm install

# Run a specific study
cd studies/01-n-plus-1-query && npm run bench:all
cd studies/02-blocking-io && npm run bench:all

• Runtime: Node.js 18+
• Language: TypeScript 5+
• Study 01: Prisma, PostgreSQL, performance.now()
• Study 02: Express, autocannon, Babel AST analysis, perf_hooks
• Study 03: Babel AST analysis, simple-git, React/Vue/Angular detectors
• Study 04: Prisma-free, pure Node.js + Python, hrtime, Babel AST
• Study 05: Prisma, PostgreSQL, EXPLAIN ANALYZE, performance.now()
• Study 06: Babel AST analysis, simple-git, discrete-event simulation, perf_hooks

• Publication: stackinsight.dev
• Tool: codeevolutionlab.com
• GitHub: github.com/liangk/empirical-study

MIT