Kincounts

Kincounts is a research project for modeling how fertility distributions shape sibling and broader kin-count distributions.

This repository includes:

• A Quarto website with methods, derivations, and analysis outputs (rendered to docs)
• An interactive Kincounts simulator (React + Vite), deployed on Vercel, with two parts: an empirical Fertility Fit view and a Simulator for kin counts
• A data folder that keeps restricted raw IPUMS extracts separate from shareable processed summaries
• Output folders with generated files

Project layout

• _quarto.yml: Quarto website configuration
• analysis: analysis pages and supporting files (including the IPUMS preprocessing methods page)
• simulator: React + Vite interactive simulator (deployed to Vercel)
• code: utility scripts, including generate_empirical_data.R, which syncs the simulator's empirical parameters from the analysis outputs
• data: raw and processed data; data/raw/ is local-only, while data/processed/ contains shareable aggregate summaries
• output: generated outputs from analyses/simulations
• docs: rendered site output
• renv-setup.R: helper script to install analysis package dependencies

Prerequisites

• R (recommended recent version)
• Quarto CLI installed and available in PATH
• Node.js 20+ (for the simulator)

Setup

Run the analysis dependency setup script:

Rscript renv-setup.R

Data

The raw IPUMS USA extract is not committed because IPUMS USA restricts redistribution of its data without permission. If you have IPUMS access, place the extract files in data/raw/. The processed aggregate summaries in data/processed/ are produced by rendering the analysis site (the preprocessing now lives inline in analysis/preprocess.qmd; the earlier code/preprocess_ipums.R script is superseded):

quarto render

The repository should include only aggregate processed files in data/processed/ so the analyses can be reproduced without redistributing record-level IPUMS data.

Build and preview the website

Render the full Quarto site:

quarto render

Rendering also regenerates the simulator's empirical-parameter module (simulator/src/lib/empiricalData.js) from the fresh model outputs, via code/build_site.R, so the simulator can never drift from the analysis.

Start live preview:

quarto preview

Rendered output is written to docs.

Run the simulator

The interactive simulator lives in simulator and is deployed to Vercel.

Local development:

cd simulator
npm install
npm run dev

Production build:

cd simulator
npm run build   # outputs to simulator/dist

Vercel is configured with the project Root Directory set to simulator (so it picks up simulator/vercel.json, runs npm install, then npm run build).

Live app: https://kincounts.vercel.app/

Notes

• Site navigation and global rendering options are configured in _quarto.yml.
• Analysis pages include both methodological derivations and simulation/model-fitting code.
• The repository may include generated artifacts in docs and output, depending on your workflow.