Practical Statistics for Analysts

What's inside

1. Level 1: Types of DataFree
   Learn to classify columns as categorical (nominal or ordinal) or numerical (discrete or continuous), and see how the type decides which summaries and charts are valid.
2. Level 2: Mean, Median and Mode and When Each MisleadsFree
   Calculate the three measures of the centre of a dataset, and see how skew, outliers and mixed groups make each one misleading.
3. Level 3: Spread: Range, Variance and Standard Deviation
   Measure how spread out values are using the range, the variance and the standard deviation, and understand what each symbol in the formulas means.
4. Level 4: Percentiles, Quartiles and the IQR
   Use percentiles, quartiles and the interquartile range to describe spread in a way that holds up against extreme values, and summarise data with a five-number summary and a box plot.
5. Level 5: Finding and Handling Outliers
   Find outliers with the 1.5 × IQR rule, then decide in a documented way whether to keep, correct or exclude each one.
6. Level 6: Charts That Tell the Truth
   Choose the right chart for each data type and avoid misleading designs such as truncated axes, 3D effects and uneven bins.
7. Level 7: The Normal Distribution, Skew and Z-Scores
   Recognise the normal distribution and skewed distributions, and use z-scores and the 68–95–99.7 rule to judge how unusual a value is.
8. Level 8: Sampling and Bias
   Learn how samples are drawn, and how selection bias, non-response bias and survivorship bias distort what a sample says about the population.
9. Level 9: The Central Limit Theorem and Standard Error
   Understand in plain words why averages of repeated samples form a roughly normal distribution, and use the standard error to measure how precise a sample mean is.
10. Level 10: Confidence Intervals
    Build and read confidence intervals for a mean and for a proportion, and describe them correctly to people who are not statisticians.
11. Level 11: Hypothesis Tests and P-Values
    Learn how a hypothesis test works, from the null and alternative hypotheses to the significance level and the p-value, and what a p-value does not tell you.
12. Level 12: T-Tests in Practice
    Compare means with one-sample, paired and two-sample t-tests, check the assumptions behind each, and report the size of the effect.
13. Level 13: Chi-Square Tests in Practice
    Test whether two categorical variables are related using a chi-square test of independence on a contingency table.
14. Level 14: A/B Test Design and Sample Size
    Plan an A/B test by choosing one primary metric, a minimum detectable effect, a significance level and statistical power, then calculate how many users are needed before the test starts.
15. Level 15: A/B Test Analysis, Early Stopping and Multiple Comparisons
    Analyse a finished A/B test correctly, and see why peeking at results, stopping early and testing many metrics or variants all increase the rate of false positives.
16. Level 16: Correlation, Causation and Simpson's Paradox
    Measure the linear relationship between two variables with the correlation coefficient, and see how confounding variables and Simpson's paradox make correlation a poor guide to cause.
17. Level 17: Simple Linear Regression
    Fit a straight line to predict one numeric variable from another, and read the slope, the intercept, R² and the residuals.
18. Level 18: Multiple Linear Regression in Outline
    Extend regression to several predictors, read each coefficient while holding the others constant, and recognise multicollinearity and overfitting.
19. Level 19: Trends and Seasonality in Time Series
    Separate the long-term trend, the repeating seasonal pattern and the leftover noise in time series data so that changes are not misread.
20. Level 20: Reporting Uncertainty to Non-Specialists
    Teaches how to tell managers and clients what a result means and how sure we are, without hiding the uncertainty or drowning them in statistics.
21. Level 21: Final Challenge: An Honest End-to-End Analysis Exam
    Use the whole course on one realistic dataset: describe it, test a hypothesis, model a relationship, check for common mistakes, and present the findings and their uncertainty clearly to non-specialists.

Access

The first 2 levels are free with a free account. Every level, the podcast edition and the AI tutor come with All Access at £4.99/month or any Creator plan — see pricing.

Start free →