Blog

Notes on building practical AI systems and statistics tools — what worked, what broke, and how to measure whether it works.

2026-10-06

Building an Offline AI That Turns Meetings Into Minutes: Six Lessons

Six lessons from building a fully offline pipeline that turns meeting audio into structured minutes: multilingual speech recognition, hallucination checks, structured extraction with a local LLM, and why practical AI is mostly engineering.

2026-10-06

Central Limit Theorem Simulation: Watch It Work (and Fail) With Interactive Demos

Watch skewed data, dice and two-humped populations turn into bell curves, see how big n really needs to be, and find out when the Central Limit Theorem fails.

2026-10-06

What Makes a Free Online Statistics Tool Reliable and Accurate?

How to tell whether a free statistics calculator is right: method choice, checks against reference software, simulation, edge cases, transparency, and the real problems our own checks caught.

2026-10-06

Interpreting Effect Size for Better Data Storytelling

Turn effect sizes into a story readers understand: Cohen's d in plain language, why the small/medium/large labels mislead, and how to report effects with uncertainty. Includes an interactive visualizer.

2026-10-06

How to Compare Groups With Chi-Square and Nonparametric Tests

Chi-square and Fisher's exact test for categories, Mann-Whitney, Wilcoxon and Kruskal-Wallis for ordinal or non-normal data: which to use, worked examples, and effect sizes.

2026-10-06

How to Document Statistical Analyses in a Report or Thesis

A checklist for writing up statistical analyses in a thesis or report, with APA-style reporting examples for t-tests, ANOVA, chi-square, correlation and Mann-Whitney, and the mistakes to avoid.