If you don't want to install Anaconda, the navigator, and

all the packages that come with them, you can instead install Miniconda.

Miniconda is a much smaller package that really only comes with conda and

its dependencies.

This is great solution for working on servers, or

even smaller machines where you just can't afford all of that overhead, and

maybe you don't have a GUI to take advantage of.

Just like with Anaconda,

you'll want to download the installer that's right for you.

Now, if you're on Windows, you'll get an executable installer.

Which means you'll just walk through a wizard to install it just like you would

most other software.

But if you're on Mac OS or

Linux, then you're gonna have to download this shell script instead.

So I'm gonna go ahead and download that.

And I'm gonna say keep, but that warning is very apt.

It can be dangerous to use shell scripts.

I'll open up the terminal and I will go into my downloads folder.

And if I look here I have this Miniconda =latest MacOSX x86_64.sh.

So to install that, I'm gonna go ahead and run that with the SH command.

And yeah, I will agree to it.

And then we have to work our way through the license.

And yes, I approve of the license terms.

They're just like they were before.

And so this asks where we want to install it.

We're gonna install this to /Users/Kenneth/miniconda.

Yours will be a different path, of course.

I'll say Enter.

And it's going to go ahead and install everything.

And then is going to ask me if I want to add the path to this to my

batch profile and I do, so I'm gonna just press Enter and so now we're good.

So I'm going to quit and

restart my terminal so that that path will be picked up.

And then, let's try this out.

So now just like we did before when we had Anaconda installed,

we can still use conda, and we can create a new one.

So this time I'm just gonna call it treehouse.

And I wanna install the requests library, no I wanna install numpy again.

Let's keep it consistent.

Now this -n is the same as doing the --name.

You can just use a shortcut if you want.

So it wants to install all these things.

I'll go ahead and say yes, and it will download and install each of these things.

Now it has to download them this time because I no longer have the Anaconda

packages available to it.

I only have miniconda installed.

So it takes it a little bit longer to install all of this stuff.

Now if you're using Miniconda, you can create your conda environments,

you can install your packages, but you won't be able to launch

the Anaconda navigator and you won't be able to use Anaconda projects.

Like if you wanted to put them onto Docker or

Anaconda cloud without installing the appropriate packages through Conda or

just downloading the entire Anaconda package.

So now I can source activate treehouse, and

I can go into Python and I can say import numpy as np, and it works.

Personally, I prefer using Miniconda to Anaconda.

That's mostly though because I don't do much, or really any data science-y stuff.

I mostly use Miniconda to easily install packages like Open CV2 because

I don't want to fight with Mac OS and Windows about drivers and libraries and

other tooling.

Miniconda saves me a lot of time and frustration.

Like I said before this is just the tip of the Anaconda iceberg.

Go out and explore the available environments, packages, guides and

tools that the Anaconda community has available.

I'm sure you'll learn a ton and find places that you can give back.

Best of luck, I'll see you next time.