How much Matplotlib you really need?

Question

I'm currently learning ML but a book which I'm referring assumes that you already know about Matplotlib, Numpy and pandas. I have somewhat knowledge of Numpy and currently working on Matplotlib more specifically pyplot as it's not really what I want to focus on I get a bit bored and distracted while learning it. So I want to know how much should I learn and where it is really used in ML. Till now I know it can be used to analyse data visually. Like, please suggest topics to learn or something like that.

Answer 1

No need to get very deep, and there are also other (arguably, better, its personal taste) alternatives for visualisation, such as seaborn. Good enough if you understand the concepts in your lessons. If you want to make more sophisticated graphs, you are going to dig into the documentation anyway.

Answer 2

Visualisation is important in machine learning for multiple reasons. When you do exploratory data analysis, for example you look for correlations between the variables, graphs can help. Also to figure out the distribution and spot any outliers.
Then at the end you might want to use some graphs to present your model and visualise the results.