juliejmurphy
commented
6 years ago Thank you for the code review.
In re-opening my project to address your review comments, I did discover that in the zip file I downloaded that there was a solution file. Certain of the notebook comments and code suggestions were different including the suggested use of "count" vs. "nunique".
-
Missing the "unique" count and instead counting all throughout the project affected many things. I have corrected that and re-ran the project from the beginning.
-
Line 206 numbers were corrected based upon use of "nunique" vs. "count".
-
Line 270 numbers were corrected based upon use of "nunique" vs. "count".
-
The sample size calculator that the Jupyter Notebook provided by Code Academy suggested is found here: https://s3.amazonaws.com/codecademy-content/courses/learn-hypothesis-testing/a_b_sample_size/index.html when I enter: conversion_rate = .15, minimum_detectable_effect = .33, level_of_significance = .9, the calculator generates a sample size of 890. When I use the sample size calculator in the solutions notebook provided by Code Academy found here: https://www.optimizely.com/sample-size-calculator/?conversion=15&effect=33&significance=90, with the same exact inputs, the calculator generates a sample size of 520. Can you advise why there is a discrepancy? For now, I have not changed my sample size.
Here are some codes that are not correctly written or don't return correct answers.
Line 58 should be species.scientific_name.nunique(), which return a result 5541
Line 206 the numbers are not correct. The right answer should be contingency = [[30, 146], [5, 73]]
Line 270 the numbers are not correct. The right answer should be contingency = [[30, 146], [75, 413]]
Line 463 the sample size 890 is not correct. Please recalculate.