Lessons
- 1Python for dataThe Python you need before NumPy and pandas: lists and dictionaries, comprehensions, reading CSV and JSON, pure functions, and grouping and sorting data.
- 2Arrays with NumPyWhy an array beats a for loop hundreds of times over, and how to read shape, dtype and broadcasting.
- 3Tables with pandasFrames, grouping, joins, and missing values you decide about instead of dropping blindly.
- 4Plot before you trainDistribution, relationship and outliers, in three plots no table of numbers will show you.
- 5Linear algebra, practicallyA vector as meaning, a dot product as similarity, a matrix as a transform, with no theorem.
- 6Probability, practicallyDistribution, mean, spread, conditional probability, and the base rate fallacy that fools everyone.
- 7Applied statisticsThe statistics an AI engineer needs: mean, median and spread, distributions, samples and confidence intervals, hypothesis testing and A/B tests, and correlation versus causation.
- 8Derivatives and gradientsWhy you step against the gradient, in one example you compute by hand and plot.
- 9Project: messy Arabic dataClean a real Arabic dataset and write a report defending every row you dropped.