Day 04 — Part 1: Statistics Basics¶
Every ML model is a statistical claim. When you say a model's accuracy is 92%, you are making a statistical statement — and without understanding distributions, variance, and probability, you cannot tell whether that number is meaningful or misleading. This session builds the statistical intuition that separates a practitioner who can explain their results from one who just runs code.
Session Overview
Difficulty: Beginner–Intermediate Reading time: ~2 hours | Exercises: ~1 hour Prerequisites: Python basics · NumPy (mean, std) (Day 02 Part 1)
What You Will Learn¶
By the end of this session you will be able to:
- Compute and interpret mean, median, and mode — and pick the right one for a given dataset
- Explain variance and standard deviation in plain English, then compute both
- Describe the normal, binomial, and Poisson distributions with real examples
- Identify when data is skewed and what that means for analysis
- Answer descriptive statistics interview questions from memory
Session Roadmap¶
| # | Topic | File | Time |
|---|---|---|---|
| 1 | Mean, Median & Mode — when each is appropriate, effect of outliers | 01-mean-median-mode.md |
25 min |
| 2 | Variance & Standard Deviation — formula intuition, np.std, ddof | 02-variance-and-std.md |
20 min |
| 3 | Probability Basics — sample space, events, independence, Bayes | 03-probability-basics.md |
25 min |
| 4 | Distributions — normal, binomial, Poisson: parameters and shapes | 04-distributions.md |
25 min |
| 5 | Cheat Sheet — quick reference formulas | 05-statistics-cheat-sheet.md |
10 min |
| 6 | Practice Questions — 10 conceptual and numerical questions | 06-practice-questions.md |
30 min |
Total active learning time: ~2 hours
How to Use This Session¶
Read files 01–04 in order. Each file has worked numerical examples — do not skip the calculation steps even if they look straightforward. The goal is not just to know the formulas but to develop number sense: what does a std of 2.3 actually tell you about spread?
Use the cheat sheet in file 05 as a quick reference during the practice questions in file 06. After completing this session, spend 5 minutes with the cheat sheet closed and try to recall the key formulas — that retrieval practice is what makes them stick for interviews.
Before You Start
Statistics is learned by doing, not reading. Have a Python notebook open alongside these notes and verify every formula with real numbers. numpy and scipy.stats are your calculators for this session.