From e85fbe784570c63f921bf521ec2e70a995891092 Mon Sep 17 00:00:00 2001 From: cfmago Date: Tue, 8 Sep 2020 21:28:16 +0100 Subject: [PATCH] adding deliverable --- lab.md | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) create mode 100644 lab.md diff --git a/lab.md b/lab.md new file mode 100644 index 0000000..d15523a --- /dev/null +++ b/lab.md @@ -0,0 +1,36 @@ + +# Lab | Reading About Statistic Concepts | Answers + +## Challenges + + +### Challenge 1: What is the difference between expected value and mean? +- The expected value is the average value of a random variable over a large number of experiments. +- The mean is the average of a set of numbers. + + +### Challenge 2: What is the "problem" in science with p-values? +- Null hypothesis significance testing is subject to certain shortcomings. +- We may misinterpret p-values above 0.05. +- We may not conclude there is ‘no difference’ or ‘no association’ when the p-value is above 0.05. + + +### Challenge 3: Applying testing to a specific case: A/B testing. + +Netflix example: + +If Netflix doesn't capture a viewer's attention within 90 seconds, the viewer will lose interest and move on to another activity. +How to allow viewers to evaluate if a piece of content is of interest to them quickly? + +Netflix knows that viewers look at the artwork first and then decide to look at additional content. +So, they focused some of their A/B testing on finding out which artwork would increase the viewership of a programe (e.g. 'The Short Game'). + +A/B test: viewers were assigned different artwork for that title. Their engagement with that title (click through rate, aggregate play duration, fraction of plays with short duration, fraction of content viewed...) was measured. + +Netflix reached the conlcusion that they could widen the audience and increase engagement by using different artwork. + +The experiment was then expanded to a wider set of titles across different popularity levels. Netflix measured the take rate (click-through rate) of all artwork variants for each title. + +The results show that Netflix raised view share of the titles testing multiple variants of the artwork and were also able to raise aggregate streaming hours. + +