Subscription and Workspaces | Azure ML Tutorial Part 2

An Azure subscription is the first step in the process of data mining in Azure Machine Learning Studio. Once we have a subscription, we can create an Azure Machine Learning Studio workspace within Azure. Azure Machine Learning is in the cloud and allows for workspaces to be shared with other users for collaborative data science projects. You can get a free trial of Azure here, and this is the link to the Azure Portal.

Hello, internet.
Welcome back to the Data Mining with Azure Machine Learning
Studio by Data Science Dojo.
Today’s video, we’re going to go ahead and create
on Azure subscription.
We’re going to create an Azure ML workspace
within that subscription.
We’re going go in and explore the features within that Azure
ML workspace.
And if you already have an Azure ML workspace,
go ahead and skip this video and go to the next video, where
I jump straight into building experiments, importing,
and exporting data.
OK, so the first thing we’re going to need to do
is we’re going to need to get an Azure free trial subscription.
The first thing you can do is you can either
type in this link.
Go ahead and pause this video and type
in this link to your browser.
Or you can just go to a search engine and then type in Azure
free trial, and it should be the first link that shows up,
So go ahead and click on that, and then it’s going to go ahead
and say Start free trial.
You will be prompted to log into a Microsoft-type account.
So that is like an email that is like @live, @outlook.
I believe Gmail is supported, so you can try that as well.
So if this is the first time you’ve ever signed up
for Azure, notice you’ll be brought to this page where
you’ll have one month that is a free trial,
and you will get $200 of credit in that first month.
So whatever one hits first, whether it be 30 days or $200.
All right, so when signing up, you’ll need two things.
You’ll need a phone number and a working phone with that phone
They’re going to text you a verification code.
The next thing you’ll need is a credit card.
So they need this to verify your identity
to make sure that you’re not a bot making subscriptions
to then create more subscriptions
to make more bots.
Another thing is they’re trying to make sure
that you don’t go from month to month
with a different email getting free Azure stuff for free.
All right, so they’re not going to charge this credit
card at all.
It’s just going to be used to verify your identity.
Now, those of you, I think, who have banks in China and India
might get charged some kind of $1 verification fee.
That is dependent on your bank, I believe.
All right, so once you have your Azure ML subscription,
to log into that subscription, you
would go to
I’m going to go ahead and paste that into my browser.
Now, you can go ahead and just go to a search engine and also
just type in Azure as well.
It would, I think,
is what you’ll be sent to.
You can just click on the Portal at the top.
I’m just going to go ahead and log in right here.
This should be a screen that you should see when you log
in to your Azure subscription.
So you’ll notice, this is our main dashboard.
This displays a bunch of tiles that is very akin to Windows.
So what we’re going to go ahead and do now is notice
that these are all the app services that we can
go ahead and spin up in the Azure Service,
but we’re here for a very specific service,
and that is Azure Machine Learning Studio.
So to make a new workspace, or to make a new anything
in Azure, go ahead and click this New button,
on the top left hand corner of your subscription page.
So go ahead click New.
You’re going to type in and search for a service
called Machine Learning, and once that is there,
you’re going to look for something
called Machine Learning Workspace by Microsoft.
So click on that, and then you’ll
get a brief description about what Azure machine learning is.
Go ahead and then click Create.
OK, so you’ll be then prompted, and so this thing over here
that just popped up is what’s called a Blade.
So in this blade, you’ll be prompted
to enter in a bunch of information
about this workspace that we’re about to create.
All right, you get to name the workspace.
So I’m going to go ahead and name it, I don’t know,
Phuc Workspace.
I’m going to go ahead and–
so if you have multiple subscriptions, which I do,
you’ll see this drop down box over here.
If you only have one subscription,
you probably won’t see this subscription here,
but it lets you basically pin this Azure
asset to that subscription to be built to.
All right, so the next thing is we need to basically pin this
to a resource group.
So you can use an existing resource group,
or you can create a brand new one.
So what a resource group is it’s a logical container
that binds cloud assets together for billing and automation
So the idea is you would pin a bunch of Azure assets
that are doing the same task, or the same job,
or for the same project, to the same thing.
Think of it like a folder, but for
your online cloud-based assets.
So that is, if you delete this resource group,
it deletes everything in the resource group.
It’s all built together, and it’s also automated together.
You can spin it all up once, if you
know how to do PowerShell scripting or things like that.
For new users, you tend to not care about what
this resource group thing is.
Just go ahead and create one.
So I’m going to call–
I normally like to name my resource group
the same thing as the asset that’s contained in it.
So I will call this workspace Resource Group.
OK, and then you’ll be prompted to, which data center
do you wish to use?
So the location of the data center
matters if you need to take in a lot of data
or if you need the output a lot of data.
So normally, you are not really charged
much for bringing in data to the cloud,
but you’re charged a lot for actually taking data out
of the cloud.
So I would recommend that wherever
you want to consume the final output of the data
is where you should go ahead and set the data center to be.
So because I want to do everything in the US,
I’m going to go ahead and select Central US,
and then it’s going to ask us to create a new storage account.
So the storage account as a separate service within Azure.
It’s called Azure Blob Storage, and what this is
is basically cloud storage.
Remember, you’re getting charged about $0.02 per gigabyte
per month to store something in here.
So this is where all your data is going to be backed up to,
and this is where all the Azure ML experiments can be saved.
Now, if you delete this storage container later,
it’s going to go ahead and, your workspace
won’t be deleted, it will just error lock,
because it no longer has the data it was referring to.
And the name of this workspace will
have to be a globally unique name, because this will become
a URL for your cloud storage.
So think of it like a domain name, like
or something like that.
So this check mark over here will tell you
that it is free and clear to be used.
It all has to be lowercase, and all
has to be in letters, no symbols, no numbers, nothing,
all lowercase and just text.
So it’s gone ahead and named it for me, so
Phuc Workspace storage.
OK, I’m fine with that.
And then, the for the pricing tier,
I think you can just leave that standard.
I think it only has standard right now.
I think they’re beta testing some other tiers, right now.
And then we’re going to go ahead and for the web service plan,
unless you really know what you’re doing,
don’t set anything here.
And basically what this is is it will set–
when you deploy web services, you’re
picking what kind of tiers of service
you want for that web service.
How robust do you want that service to be?
How many people and transactions do you want it to support?
For the most part, the free one is fine,
where we have 1,000 transactions is fine.
So we’ll create a brand new web service plan for that,
and we’ll select that as a tier, which is no pricing.
We’ll select the standard tier, which charges us nothing,
but we only get 1,000 API calls.
Which I think is more than fine, especially
if we’re just prototyping.
OK, so once we filled everything out here,
I think we’re good to go, we will
want to pin this asset to our dashboard.
So remember those tiles we saw at the beginning?
That’s where we want our asset to be,
so we can always refer to it later.
So go ahead and click the Create button now.
So this will take about two minutes to create.
Go ahead and do something else for those two minutes.
You will see a tile that has now appeared, because we selected
that button.
It says Pinned to the Dashboard, and this
is going to spin for the next two minutes.
All right, it looks like it’s finished creating,
and it automatically brought me into the asset.
But if you don’t know how to get back here,
if you’re on the dashboard, you can just
click on the tile that was pinned.
So click on the tile and you’ll get back to this page.
So this page lets you basically manage the Azure asset that
is the Azure ML workspace.
And now to get to the workspace itself,
you will click this button under Additional Links that says
Launch Machine Learning Studio.
Now, what I actually prefer is this actually
goes through a separate website.
So you can actually go to Azure ML
by just going to
That does the same thing.
So if you take this URL and paste it into your browser,
it’ll take you to the same place as clicking this button.
So I actually prefer this URL, because it cuts out
the middle man.
Because the idea is, all right, I have to log into Azure,
and then once I’m in Azure, I have to find my asset.
I have to click on my asset.
Once I click on my asset, I have to click on this Launch Machine
Learning Studio.
Or, if I just want to use Azure ML,
I will just go directly to this URL,
and I’ll cut out Azure altogether.
I’ll go directly into Azure ML Learning Studio.
That’s just a tip that I’ll give to you.
All right, and it’s going to ask you to sign in again.
It will share your Azure subscription.
So that’s fine.
So go ahead and log into your Azure subscription,
but this time do so by logging into your Azure subscription.
So notice, I’m inside of Phuc workspace.
So that is the name of my workspace.
So notice that you can have lots of different workspaces,
and notice that you can select different regions
and things like that.
So there’s nothing stopping you from having lots of workspaces.
So you can also change workspaces up here.
So I notice I don’t have anything in West Central US,
but I think if I go to South Central US,
I have four other workstations I can select from.
So notice, these are all self-contained workstations
that are either been, A, shared with me,
or I’m hosting them on a separate account somewhere.
So let’s go back to our current one.
So that’s how you switch workspaces.
Now, the reason for this is if you go to, for example,
the Setting button over here, you
can invite users to your workspace.
So that is, if you have a team, if you’re working on a project,
you make a new workspace and invite all your team members
to that workspace.
And that will be a self-contained workspace.
So this is the closest thing that we data scientists
have to a Google doc right now.
This is probably one of the coolest collaboration
tools we’ve had in a while.
So you can always do this.
You can invite more users, for example,
and then I can invite my buddy Eric at,
for example, and I can go ahead and make him a user.
So now, Eric can go ahead and log in
to see the same experiment, data models, everything
that I would see in this together.
And then, notice that if I want to switch teams or projects,
I would switch the workspace.
Remember, you’re charged $9.99 per workspace,
so keep that in mind.
So the data sets that you will bring
into Azure will be saved here.
They are basically objects.
Any models that you train will be in here.
There’s also a new feature, which is Notebooks.
So a brand new way of programming
that’s taking the programming world by storm is Notebooks.
The idea of a self-contained environment
that can be deployed to the web, where you just code on the web,
and then you can use almost any language that you
would want to.
And you can share those notebooks
or just expose them as web services, which is really cool.
So right now it supports Python and R Notebooks.
So those are the primary programming languages
right now in the open source world for data scientists.
If you have any deployed web services, they would be here,
but your bread and butter will be this guy right here,
the Experiments tab.
So an experiment is what they call a file, for data science,
instead of Azure ML.
So just like a spreadsheet.
A spreadsheet file in Excel is called a spreadsheet.
And a Word document called a document.
These are called experiments, and then you
can pin multiple assets together into what’s called Projects.
And notice, I can Create a Project, name the project,
and then I can pin various experiments, web services,
and data assets to this project.
So notice I can have multiple projects going
on at the same time for the same team,
and everything will be good.
And how I got here was, see this thing in the top
left hand corner here, you can click here
and click on the Studio.
So Azure ML actually has three other pages associated with it.
For the most part, you almost always
want to be in this Studio Mode, right here.
And I think we’re out a time for this video.
Go ahead in watch the next video,
where I will go ahead and show you
how to create your first experiment,
import data, export data.
And if you want to see more videos like this in the future,
go ahead and like and subscribe, and I will look forward
to seeing you at our boot camp.
See you next time.

Part 3:
Import and Export Data, Modules, and Experiments

Part 1:
What is Azure Machine Learning?

Complete Series:
Introduction to Azure Machine Learning

More Data Science Learning Material:
[Video] Intro to Clustering: An Unsupervised Learning Technique
[Blog] The Best Data Science Podcasts
[Blog] Custom R Models in Azure Machine Learning Studio


Phuc H Duong
About The Author
- Phuc holds a Bachelors degree in Business with a focus on Information Systems and Accounting from the University of Washington.


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>