Age-based counting functions in a microsimulation are easy to write as nested loops and brutally slow at scale. This tutorial rebuilds them with a difference-array and cumulative-sum kernel — turning every range update into two events and one cumsum() — and benchmarks the speedup live in R.

If you use Stata for analysis but are curious about R, this is a side-by-side guide to the equivalents — working directories, packages, data loading, exploration, missing data, renaming and labelling variables, merges, IDs, frequencies, and tabulation.

Stop relying on web converters or paid PDF tools. Use Python with pdfplumber and tabulate to extract clean text and structured tables from PDF documents — useful for data wrangling, NLP preprocessing, and any workflow where the source data lives in a PDF.

A hands-on COMPAS case study using Aequitas for group-level fairness metrics and AIF360 for reweighing-based bias mitigation. Covers demographic exploration, false-positive disparity by race, and pre-/post-processing mitigation with native Python chunks.

A hands-on introduction to quanteda built around an unlikely teaching corpus: every line of dialogue from How I Met Your Mother. Covers corpus construction, preprocessing, document-feature matrices, similarity, networks, and collocations.

data.table on macOS ships single-threaded by default — Apple’s bundled clang has no OpenMP. This guide walks through installing LLVM via Homebrew, configuring ~/.R/Makevars, reinstalling data.table from source, and verifying the multi-core build on both Intel and Apple Silicon machines.