
Another reason that visualizing can be very helpful is there's

some data that just isn't going to be very good.

There's just going to be some data that's going to be difficult to fit

with a regression.

Let me show you what I mean with a quiz.

Here are a few different data sets I might have.

And I'm making a scatter plot of some features of each one.

And my question for you is, which of these data sets would you look at it and

say, that looks like a good candidate for linear regression?

We've done a lot of examples of linear regression, and by a good candidate,

I mean you can write an equation y equals mx plus b

that does a good job of describing the patterns in the data.

Place a check box by all the ones that would be good candidates.