No problem, AFAIK data analysis & machine learning enjoy are easier/perform better with more data, so i thought obtain past data once might be useful in your case.
Yeah, For now I'm getting all the data possible from my local client, than I'll se whether to integrate something I can't get already there like mempool size, if I'll be not happy with te results I'll get
This tool can be useful in your case, it parses the blockchain and get useful information from it. But it does it without using a wallet, it takes the binaries files of the blockchain directly.
Check it here: https://github.com/NicolasChoukroun/blockparser
It is in C and can be compiled easily on an unix OS.
Awesome thx for point me to this tool