Tidying a BLS table whose two-level headers spread five measures across two years, then comparing state unemployment rates and their change from 2023 to 2024.

Tidying a Wikipedia electricity table with multi-level headers, then asking which countries are most sustainable and how they group by dominant energy source.

Tidying the World Bank life expectancy table from wide to long with tidyr, then comparing countries over time and across the COVID-19 period.

Calculating each player's Elo expected score from the Project 1 chess tournament, and listing the biggest over- and underperformers.

Tidying a small airline delay table in R, then comparing two airlines overall and city by city — a worked example of Simpson's paradox.

Parsing a formatted chess tournament text file in R into a tidy CSV with each player's name, state, points, pre-rating and average opponent pre-rating.

Year-to-date and six-day moving averages of Apple, Microsoft and Amazon closing prices, calculated with SQL window functions and loaded into R.

A Global Baseline Estimate movie recommender in R, checked against the course spreadsheet and applied to my Week 2 movie ratings.

Null error rate, confusion matrices, and accuracy, precision, recall and F1 at thresholds 0.2, 0.5 and 0.8 for a penguin sex classifier, with use cases for each threshold.

Six friends rate six 2025 movies: a normalized SQLite database loaded into R, with a tested strategy for missing ratings and a check on whether standardizing ratings helps.

Loading 25 seasons of EPL match data into R, selecting and renaming columns, decoding the full-time result target variable, and checking a COVID-era home-advantage study against the cleaned data.