I am soon releasing a liveProject with Manning on doing data science with Julia. You can find the link to it here.
The liveProject is - as its name says - project oriented, so expect it to be challenging for someone just starting to use our ecosystem. I assess it as intermediate level of complexity. It is best suited for someone who already has some introductory level knowledge about JuliaData ecosystem.
The tutorial consists of 5 mini-projects and the first of them will be available to go though it for free.
Looks interesting, although I wonder why you didn’t consider using MLJ for the final bit (which wraps XGBoost.jl of course, but also offers loads more opportunities)?
This is a live project, so users are free to use any ecosystem they want prepare the solution. We did not go for MLJ.jl as it turned out that (unfortunately) there would be too much to explain to the user before the model we want to build could be built. Still we mention that MLJ.jl could be used.
Essentially my conclusion was that MLJ.jl deserves a separate liveProject in order to explain all concepts that are needed to be learned so that the reader can reliably use it.