As an older scientist interested in data science, I find your videos (and associated webpages) particularly useful. They are always well laid out and easy to follow. I feel bad for not commenting sooner since I have watched many of your videos over the last several months. Figured it was time I let you know how much I appreciate them. Thank you!
Fantastic! I hope you'll do more videos in the future covering useful techniques for tackling inference problems. For data scientists, I think it is equally as important of a topic as predictive modelling.
Great presentation. I ran into this package on datacamp, and it's great to get some additional information about it. I do like the "pipeline" approach and some of the handy visualization routines that are part of the package as well. I'm definitely looking forward to seeing more your content. Looks like you have a lot of great topics here. Thanks for putting it all together!
Thanks Julia. This was more than helpful. I particularly like how the function in R has made bootstrap simulation and computing chi square just a one line code. I was also hoping, one day you will make a video about how to communicate model findings to non technical audience, maybe?
Great job Julia. We use Chi Square method to test if two categorical variables are associated with each other or are independent. That's why it is also known as test for independence.
HA yes certainly that is right (if very basic info), but that doesn't change that the chi-squared statistic (not an everyday number) can be hard to have real intuition about. You might want to be careful in the future about assuming that someone doesn't know very introductory definitions like this.
hey, love your work as always, you're awesome. i've seen that you have done a PCA analysis exemple, can you do an FCA (factorial correspondence analysis) exemple, i'm curious to see how it's done by the tidymodel library.
And for anyone new to this field: The book 'Statistical Inference via Data Science: A ModernDive into R and the Tidyverse' provides a great introduction to not only key concepts of statistical inference, but also the infer package which Julia used in this video!
Best channel on YT. Thank you. Also, do you have any suggestions for handling “big p, little n” data (i.e more predictors than observations)? Is “deep learning” appropriate in this situation?
No unfortunately, deep learning is almost certainly _not_ a good approach if you don't have a lot of observations (although there are tricks with using pre-trained models in certain domains). I suggest you check out Ch 18 of ELSR for some ideas.
I didn't have much formal stats training from my physics degree and PhD (only a data analysis class in grad school!) so I always feel like I am playing catch-up and continuing to learn! One book that I like a lot for statistical learning is ISLR: www.statlearning.com/
On a Mac, it is Cmd+Option+I. You can see lots of shortcuts here (or in the IDE under "Tools"): support.rstudio.com/hc/en-us/articles/200711853-Keyboard-Shortcuts-in-the-RStudio-IDE