I like that it's open to people who might be new, or have little knowledge of data-science/ML but can still use C++ for programming purposes with ease (and without having all those libraries).
If you are already familiar wth Python & Pandas then this won't help much as there aren' t many tools in python which will allow one access what is needed directly from CLI commands - so most things need tinkering around using code first before being able do something useful! This allows me quick results if i know exactly where my problem lies within an image(which may not always give clear details) even when working via command line interface only rather than visualizing through GUI etc..I'm solving problems relating mainly into object recognition such issues related specifically towards medical images / radiology reports eectually any kind off "big" dataset really! The benefits come mostly outof just knowing how big your datasets actuallyis by learning about some common mistakes made.