Stage 2 of 10 · 8 lessons · ~6 weeks

Statistics and probability you actually use

Describe data honestly, understand distributions and uncertainty, and learn which test answers which question, so you can tell a real pattern from noise.

Start: Describing data: centre and spread

Lessons· one lesson for each topic, in this order

Do this

Describe three columns of a real dataset with the right centre and spread for their shape.
Plot a histogram of each and name the distribution it resembles.
Work one base-rate problem with counts for 10,000 cases.
Simulate 1,000 samples in code and plot the sample means.
Compute a confidence interval for a mean and for a proportion.
Pick and run the right test for two questions, and write what each result means.

Start with· follow one of these

VideoFree
StatQuest with Josh Starmer
StatQuest · Beginner
Free: short, clear videos on each concept, from the normal distribution to p-values and power
CourseFree
Khan Academy Statistics and Probability
Khan Academy · Beginner
Free: the probability, sampling distributions, confidence intervals and significance tests units, with practice

Practise

OpenIntro Statistics
Free textbook with end-of-chapter exercises on real data, and labs in R
Think Stats
Free: statistics taught through Python code on real datasets, with notebooks you can run

Keep handy

statsmodels Documentation
The statistics and stats tests pages, when you need a test SciPy does not have

Also worth your time

Seeing Theory: A visual introduction to probability and statistics (opens in a new tab)Read
Interactive visuals for probability, distributions, inference and regression: play with each one
Mathematics for Machine Learning and Data Science SpecializationGo deeperPaid
The probability and statistics course in this specialisation goes further, ahead of heavier modelling

Before moving on: A notebook on a real dataset that describes three columns honestly, computes two confidence intervals, runs two tests you chose from the question, and explains each result in two plain sentences a manager would understand.

Tauseef's advice

You do not need a statistics degree, but you do need to explain what your numbers mean. After each concept, explain it out loud without jargon; if you cannot, watch the StatQuest video on it and try again. Simulate: drawing random samples in code and watching the averages settle teaches the central limit theorem better than any proof.