Lack of Standard Analytics Pipeline
Hello all,
I’m quite confused (and probably naive) as to why there isn’t a seriously structured & comprehensive pipeline format that most/all data analysts use when selecting/executing their potential models.
Imagine a world where you upload your data set to some sort of entity. You answer a few preliminary questions (ie. I care about explainability, your business objective is xyz, etc.), to where you get pipelined to the next unique step given your previous answers. Maybe some of your previous answers implies that you should then clean the data up this way/do this to the data. Then, given the way you cleaned your data/your goal/your output variable parameters, you’d be suggested to use “business knowledge” or “apply parameters”, or be prompted to do a preliminary heterosekastic analysis, etc.
Idk. I’m finishing up my Analytics Masters’, and feel like I’m constantly told that this isn’t probable since every question is unique + you need domain experience, but it seems that no matter what projects I work on, there’s always similar steps I do. Idk.