11 Comments
User's avatar
Neo's avatar

Is the modelling part really that simple with autoscktlearn?

Oh man I have really been wasting my time trying to learn the modelling aspect.

Only upside is that I’ve been focusing on steps 1, 2 and 3.

Would there actually be any value in learning the theoretical components behind each of the modelling techniques you have mentioned though?

I know that one of the data scientists has used some k nearest neighbours clustering analysis based on our datasets.

Expand full comment
BowTied_Raptor's avatar

Watch this video, this guy knows what he's talking about, and is one of the few good sources of information out here.

https://www.youtube.com/watch?v=5gBi8p42HX4

Completely answers your question as well.

Expand full comment
Sean's avatar

That guy kinda reminds me of you.

Is he BowTied_Raptor in disguise??? 😂

Don’t worry, you don’t have to reveal your secret. 😉

Anyhow, I was watching his videos and he mentions that if you’re not in a data position already to pretty much give up on getting any role outside of a data analyst role. Do you agree with him on that?

Expand full comment
BowTied_Raptor's avatar

Real recognizes Real. Now you know why I recommend him.

I also recommend everyone start off with some sort of a data role. Could be data science intern, or could be a full time data analyst.

Data Scientist/Data Enginer/Machine Learning engineers are all mid level roles. ALL job postings say you need atleast 3 yrs exp.

Read my post on $300k in 3 years. I started off with a shitty $45k/yr data analyst role. We all have to suck it up in our starting points, then we can hop to the real ones.

Expand full comment
BowTied_Raptor's avatar

Also, his full name is: Mike West.

Here's his quora, grab some popcorn, and sit back and read some of his answers.

https://www.quora.com/profile/Mike-West-99

This guy's got the patience of a god, lmao.

Expand full comment
Defi Dagoth (∞)'s avatar

Hey, do you know if his "logikbot" course is any good?

Expand full comment
BowTied_Raptor's avatar

If your python skills are a bit weak, I'd give this one a shot: https://www.logikbot.com/courses/data-cleansing-master-class-in-python

If you want to upgrade your SQL skills: https://www.logikbot.com/courses/the-fundamentals-of-transact-sql-for-machine-learning-engineers

But the SQL stuff is a bit more on the advanced side than what I use daily. Still useful though.

Expand full comment
BowTied_Raptor's avatar

Is modelling really that simple?

-> Yes

Having I been wasting time trying to learn the modelling aspect

-> If that's the only thing you focused on, then yes.

Would there actually be any value in learning the theoretical components behind...

-> Only for the technical interviews, but outside of that, nope.

KNN cluster analysis is just a simple algorithm that's used to predict a certain y value. KNN only works if the data points themselves are clustered. KNN is also frequently used to fill in missing data points by using its surrounding neighbors, I will make a post on how this is done, soon.

Expand full comment
BowTiedCelt's avatar

Captains of crush during meetings another good challenge

Expand full comment
BowTied_Raptor's avatar

oh yea, the forearm trainer.

Get them big elbows.

Expand full comment