Bravo! Love the fact that you took the time to show what an outlier might look like so as to change ones interpretation: the visual adds SO much to the lesson.
Greetings and best wishes to my youtube teacher @Dataslice! caught you after a long time. Thank you for another great video. A request followed... please make a series of videos on Regression covering its types, usage according to situations.
Hey, great video as always! Just a question: whats the benefit of looking at the graphs to check for homoscedasticity/normality rather than performing tests like Shapiro-Wilk or Breusch-Pagan? I'd also like to see a video about most common/useful linear transformations in the future (just a humble request). I always learn something from your videos, keep up the great work!
Great question! Using Shapiro-Wilk/Breusch-Pagan is a convenient way to quantify normality/homoscedasticity since they give you a p-value to interpret, but the graphs give you a more qualitative interpretation so you can understand normality/homoscedasticity in the context of your dataset. I agree that there may not be a huge added benefit of looking at the graphs, but in general, the combination of quantitative statistics + the qualitative interpretation of the plots can give a clearer understanding of the linear model. So tl;dr, I think it’s best to understand all the possible tools/tests/graphs available and use them as needed! Linear transformations would be a good topic to cover - I’ll add it to the backlog! :)