dspiegel29 / artofstatistics Goto Github PK

This is a wonderful book.

It is 0.000651 in the R file. So, its inverse turns out to be 1535 days approximately as written in the book.

Maybe 0.065% would have been more appropriate for 1535 than 0.07% because 1/0.0007 = 1429 and 1/0.00065 = 1538, which is much closer to 1535.

As a data scientist who mainly works and codes in Python, I would love to see the code examples available in Python, too. I will try to provide them by myself, so right now I am adding Jupiter notebooks to the folders which will produce roughly the same analysis an pictures as the R examples. Once I am finished with this, we might have to think about how to structure the code and output because right now the final output is given as html, and I don't want to provide two html files. Maybe we will group the code in sub folders according to the programming language.

The aim will be to roughly reproduce the pictures so the reader has a starting point, but I will not recreate the exact picture because this will be too much the effort. We should follow the Pareto principle here. So for example the first graphic I reproduced with plotly and their margin functionality which creates the histograms but with other bin sizes.

But first I will have to migrate all examples, and I have just started reading so I am looking forward to this!

The book says

" ... so relative to the true answer the median was an over estimate by around 10%, and only around 1 in 10 people got that close."

The median is about the number of jelly beans, not the number of people. The number of people who guessed within the range of true value plus/minus 10% will be a lot more than 1 in 10.