Is there a way to automatically compile a list of questions/answers ranked from easy to hard that would form a tutorial for learning Statistics from scratch?
Currently SE is a Q&A site, but it would be more profitable if you could turn the Q/A into course materials assuming every question about the subject has been answered.
Profitable for whom? The question is germane as Stack Exchange is a business, although most of us pay nothing and get no income either.
If you mean useful for students, then surely, but that is just not the main way that SE (including CV) was designed or is maintained. CV for example is messier than Wikipedia (although on many topics more nearly authoritative) and there is much overlap and even much material that individuals may think is badly written, confused or incorrect. Almost all threads arise in response to particular questions and there is no overall design to cover all possible topics or even to identify and fill in gaps. There is no managing committee working to produce a kind of statistical encyclopedia or series of online textbooks.
I really wouldn't assume that every question has been answered; there are different new questions every day and statistical science isn't static either. Learners at any level can search using keywords that are of interest or the names of contributors whose style is congenial and who have obvious expertise.
I would see the main strength of CV as going beyond basic texts or courses in various ways, including
Questions that are often glossed over, or not well explained, in such places, possibly because the point is both too arcane for elementary treatments and too trivial or mundane for advanced treatments (really good discussions of when and how to use transformations or how or even whether to worry about outliers seem quite unusual)
Questions that arise because many treatments are muddled, incorrect or contradictory (some spectacularly popular texts are riddled with error and confusion)
Questions which involve experience, judgement and personal analysis or modelling style in a way that seems open to discussion (many, perhaps most, questions based on the OP's own dataset can be of this form)
A corollary is that really elementary questions may not attract good threads here, possibly because too few people are interested in writing good answers, or are dubious that they can improve on what is readily available if you look for it.
Those most active in answering questions here are by definition sympathetic to posting explanations on the internet, but (I speculate...) almost never did they themselves learn statistics or machine learning mostly by Googling and watching videos. They attended courses and (perhaps especially) worked their way through textbooks, studied papers, and learned by analysing data and making mistakes along the way.
