type: Chore
Academic project: We have a new version for a webpage and need to find out what layout would like best: old vs new.
To do so, they run an experiment with a Control Group receiving the old page and Experiment Group receiving the new one and they count the number of conversion within each group.
We have the population data, but we run also sampling distribution to find deeper statistics. For sampling distribution we use np.random.choice and np.random.binomial, find p-value and discuss its meaning. And find the same conclussions using statsmodel Logistic Regression to discard other possible data variables such as country.