Joking aside, this is an important point:

If you're a beginner, you'll probably get a *lot* more use out of linear regression, logistic regression, and decision trees.

[thread]

#data #DataScience #MachineLearning https://twitter.com/tunguz/status/1328725883784409089

2/

If you're a relative beginner in data science, and you're focusing on deep learning, you're getting way ahead of yourself.

#data #DataScience #MachineLearning

3/

Deep learning is really cool, but in many businesses that need data work, it's uncommon.

Deep learning is an advanced topic with niche applications.

#data #DataScience #MachineLearning

4/

Meanwhile, linear regression, logistic regression, and decision trees still do a lot of heavy lifting in analytics organizations.

#data #DataScience #MachineLearning

5/

Why?

Linear regression, logistic regression, and decision trees are:
– easy to learn
– easy to use, and
– easy to explain

#data #DataScience #MachineLearning

6/

Moreover, for a few legal and regulatory reasons, they're also preferred over advanced "black box" models for many use cases.

7/

Not to mention ...

Linear regression, logistic regression, and decision trees are still very powerful.

They are still very strong tools that get the job done for a variety of use cases.

#data #DataScience #MachineLearning

8/

I can't stress this enough ...

If you're a beginner, forget about deep learning and the "cool" techniques.

#data #DataScience #MachineLearning

9/

Focus on foundational data science techniques like data manipulation, data visualization, and data analysis.

And for machine learning ...

Master traditional techniques first, like linear regression, logistic regression, and decision trees.

#DataScience #MachineLearning