The Art of A/B Testing

The Art of A/B Testing

0%

1. What is A/B testing?

Welcome to the A/B testing course. In this section, we will tell you what controlled experiments are, where they are used, and why they are used. We will look at a specific example of a controlled experiment and talk about A/A testing. The Python programming language will help us with this. Well, let's get started!

Prepare for the Test

2. Normality Check

Before conducting a controlled experiment, we need to determine the characteristics of the sample. The type of tests that we will conduct a little later will depend on these characteristics. You will learn to build distribution histograms, box plots, and check the normality of the distribution. Let's get started right now!

About Normality

Descriptive Statistics

Challenge: Descriptive Stats

Histograms and Box Plots

Challenge: Plotting Hists

Challenge: Plotting Boxplots

Challenge: Shapiro Test

3. Variances in A/B Testing

The last thing to do before conducting a controlled experiment is to check the sampling variances. Some types of graphs and Levene's test will help us with this. Let's talk about it in more detail in this section!

Violin and Swarm Plots

Challenge: Violin and Swarm Plots (Part 1)

Challenge: Violin and Swarm Plots (Part 2)

Challenge: First Levene's Test

Challenge: Second Levene's Test

4. T-Test

So we got to A/B testing. Soon you will make your first parametric tests, or as they are also called - T-tests. We'll use all the knowledge we've learned in the previous sections: metric definition, normality testing, variance testing, and plotting. All this will help us get statistically significant results. Well, to work!

Graphical Results of Previous Studies

The First T-Test

Challenge: Second T-test

The Third T-Test

Challenge: Fourth T-Test

5. U-Test

It turns out that sometimes the samples do not have a normal distribution. This means that we cannot perform parametric testing. What to do in such a situation? Non-parametric testing. We will talk about him in the last section.

Challenge: Define metric

Challenge: Another Normality Test

The First U-Test

Challenge: The Second U-Test

Challenge: The Last U-Test

The Third U-test

Course Content

The Art of A/B Testing

Shapiro Test

The Shapiro Test is a statistical test that is used to test the hypothesis of a normal distribution. It compares the distribution of the data with a normal distribution.

The null hypothesis assumes that the data are normally distributed. If the p-value is below the significance level (below 0.05), then the null hypothesis is rejected.

In such a case, we can argue that the data is not normally distributed (the alternative hypothesis is accepted).

Let's run the Shapiro Test for the first columns from the control and test groups at the same time:

Great! We got two results.

The higher the Statistic value, the more evidence is found in favor of a normal distribution. The p-value in both groups is high (greater than 0.05), which means we accept the null hypothesis.

Both columns are normally distributed.

Note
If we have more than 5 000 observations, it is better to use the Kolmogorov-Smirnov test. Its use is similar to the Shapiro test.

Everything was clear?

Section 2. Chapter 7

some-alt