A browser-based modular canvas for learning statistics by building it: wire together sampling, estimation, hypothesis tests, the bootstrap, and diagnostics, and watch each result update live. Every patch is reproducible from a seed, and the numerics are verified against R.

Interactive tool for exploring how family background (parental education, occupation, and wealth) relates to children's educational achievement at age 15 across 100+ countries using OECD PISA data (2000–2022). Compute survey-weighted statistics, ESCS gradients, and cross-country comparisons.

An R package that turns PISA, TIMSS and PIRLS data into the quantities education researchers report — socio-economic achievement gradients, school segregation indices, and the Ferreira–Gignoux inequality of educational opportunity — with correct plausible-value pooling and replicate-weight inference. It also decomposes these quantities, compares them across countries, and runs specification-curve (multiverse) analyses, and is cross-validated against the intsvy and segregation packages.

A multiverse / specification-curve lab on a real longitudinal wage panel (NLSY, 545 men, 1980–1987). Pick an effect — the union, marriage or schooling premium — and see the whole universe of estimates open up as you change estimator (Pooled OLS, Between, Random Effects, Fixed Effects, First Differences), controls, sample and inference. Every estimate is verified against R's plm.