43 ms·

I'm continuing work on the No Bullshit Guide to Statistics, which is an introductory book on statistics that includes all necessary prerequisites (practical dat

I'm continuing work on the No Bullshit Guide to Statistics, which is an introductory book on statistics that includes all necessary prerequisites (practical data analysis using Pandas and probability theory math using the computer models from scipy.stats). It's been a many-year project (because STATS IS TOUGH!), but I've had very good progress in the last couple of months of 2022 (finished first few chapters) and now keeping the momentum going into 2023. It's really hard to explain statistical concepts without bullshit... there are all kinds of formulas and rules of thumb, so it sometimes takes me 2x or 3x rewrites before I find an explanation that is simple enough and free of unnecessary historical baggage. One thing that I have going for me is the ability to use computer simulations to check all the equations, and generate nice visualizations. Seeing the equations + graphs + code in parallel makes stats concepts finally click... at least for the author, I hope for the readers too ;) If you're interested in checking this out, I maintain a continuously-updated book outline here: https://docs.google.com/document/d/1fwep23-95U-w1QMPU31nOvUnUXE2X3s_Dbk5JuLlKAY/edit# https://docs.google.com/document/d/1fwep23-95U-w1QMPU31nOvUn... I also have a mailing list where I send chapter preview PDFs as they become ready (see the end of the gdoc).

dinkleberg 4y ago
This sounds excellent. Will be keeping an eye on this. As I've been spending more time doing data viz recently, I've started to realize how little I remember about stats from college and am having to study back up. Unfortunately, the resources I've found have left much to be desired. My best bet at the moment seems to be going through MOOC courses to essentially retake my stats courses. But your guides here seem much nicer.
- ivansavz 4y ago
  > Unfortunately, the resources I've found have left much to be desired. Yeah most textbooks out there are pretty bad (aim to cover the standard curriculum, which is not very useful for most practical statistical analysis tasks). Recently however, there has been a growing number of teachers/educators adopting the "modern statistics curriculum" based on resampling methods (bootstrap estimation and permutation tests), so you should look for those if you want a review that is more friendly to learners with a coding background. I have another google doc where I've saved the best links to learning resources, you might want to check out while waiting for the book (it might take a while!): https://docs.google.com/document/d/1GWyuNC4qCL5ecYSoY6GxvuhOz4AxYEhIKzSS_lgEGaY/edit# https://docs.google.com/document/d/1GWyuNC4qCL5ecYSoY6GxvuhO... In particular the talks by Vanderplas and Downey in the section "Tutorials > Resampling methods" are all excellent. Also, here is a four-part video tutorial + notebooks that I prepared, which gives an overview of all parts of modern curriculum (from data management, through prob theory, to stats modelling): https://nobsstats.com/stats_overview/README.html https://nobsstats.com/stats_overview/README.html

dinkleberg 4y ago

This sounds excellent. Will be keeping an eye on this. As I've been spending more time doing data viz recently, I've started to realize how little I remember about stats from college and am having to study back up. Unfortunately, the resources I've found have left much to be desired. My best bet at the moment seems to be going through MOOC courses to essentially retake my stats courses. But your guides here seem much nicer.

ivansavz 4y ago
> Unfortunately, the resources I've found have left much to be desired. Yeah most textbooks out there are pretty bad (aim to cover the standard curriculum, which is not very useful for most practical statistical analysis tasks). Recently however, there has been a growing number of teachers/educators adopting the "modern statistics curriculum" based on resampling methods (bootstrap estimation and permutation tests), so you should look for those if you want a review that is more friendly to learners with a coding background. I have another google doc where I've saved the best links to learning resources, you might want to check out while waiting for the book (it might take a while!): https://docs.google.com/document/d/1GWyuNC4qCL5ecYSoY6GxvuhOz4AxYEhIKzSS_lgEGaY/edit# https://docs.google.com/document/d/1GWyuNC4qCL5ecYSoY6GxvuhO... In particular the talks by Vanderplas and Downey in the section "Tutorials > Resampling methods" are all excellent. Also, here is a four-part video tutorial + notebooks that I prepared, which gives an overview of all parts of modern curriculum (from data management, through prob theory, to stats modelling): https://nobsstats.com/stats_overview/README.html https://nobsstats.com/stats_overview/README.html