Complete guide to hands-on A/B Testing | A/B testing in Python | All that you need to know
Вставка
- Опубліковано 30 вер 2024
- 🔬 In this video, we cover hands-on demonstration of A/B testing using real-world data from Kaggle! We'll guide you through a comprehensive exploration of A/B testing, using a dataset that examines the conversion status when an ad is displayed to a test group versus a general public service announcement.
📊 Our journey begins with a meticulous exploration of the dataset, which includes variables such as the day of the week with the highest ad displays, the hour of the day with the most ad displays, and the total number of ads shown to prospects. We'll methodically pair each variable with the conversion status, creating insightful visualizations like stacked bar charts, pie charts, and box plots to uncover meaningful patterns and trends.
🔍 With our exploratory analysis complete, we'll move on to the heart of A/B testing-statistical hypothesis testing. We'll perform a proper chi-squared test of dependence to assess the relationship between categorical variables and the conversion status. Additionally, we'll conduct a Mann-Whitney U test to compare the distributions of a continuous variable between the two groups, providing robust statistical validation to our findings.
🚀 This video will help you master the art and science of A/B testing as we bridge the gap between theory and practice, empowering you to leverage data-driven insights for impactful decision-making ! 📊
Happy Learning!
You explain things so well and in so details.. hello i am a fresher data analyst... is there a way to connect you??
Thank you! Your encouragement means a lot. We have some thoughts in the pipeline for better connect with our patrons like you. Probably we'll be able to get back soon.
Same with me. I would really like to connect to you@@prosmartanalytics
Your way of teaching is very clear and understandable, is there any way that we can connect?
Thank you! Sometime in near future, till then please stay connected with this channel.
you guys have a website?
Yes, it is sixsigmaprosmart.com
thanks a lot for this great video; is it possible to access the notebook code as well?
Thank you! Due to some IP infringement issues in the past, we don't post the code.
Great video! But may I ask why we wanted to use t-test and not z-test? given we are dealing with a large dataset? Thanks!
Good question! The biggest challenge associated with z test is that it requires the population standard deviation. Knowing population standard deviation means we know the population, and if we know the population we won't need inferential statistics. Therefore, for all practical reasons generally a t test is preferred. 😊
@@prosmartanalytics Got it!! Thanks for the prompt response!!!
Welcome! We have a complete playlist on hypothesis testing, you may go through it if you are interested in this topic.
@@prosmartanalytics hello. but can the standard deviation not be calculated?
@@taibatabowaba9828 We can easily calculate sample standard deviation, but knowing population standard deviation is not so easy. To be able to calculate population standard deviation we first need to have access to the population. If we have access to the population there won't be a need to perform inferential statistics and hypoyhesis testing.