Subscriptions & Workspaces | Intro to Azure ML Part 2

In this informative video, the presenter guides viewers through the process of setting up an Azure Machine Learning (ML) workspace. Starting with obtaining a free trial subscription from Azure, the tutorial covers logging in with a Microsoft account, verifying identity with a phone number and credit card, and navigating the Azure portal. The presenter then demonstrates creating a new ML workspace, explaining the significance of a globally unique workspace name and the resource group concept. The video also touches on data center location considerations for data input and output costs, the creation of a storage account for data backup and experiment saving, and the selection of a web service plan. The presenter concludes with tips on accessing the workspace directly through Studio.AzureML.net, inviting team members to collaborate, and managing multiple workspaces and projects within Azure ML Studio.

Takeaways

• 🌐 Start by creating an Azure subscription and an Azure ML workspace within that subscription.
• 🔍 Use the Azure free trial by signing up with a Microsoft or Gmail account and providing a phone number and credit card for identity verification.
• 💳 The credit card is used for identity verification and will not be charged unless you exceed the free trial limits.
• 📌 Log into your Azure subscription at Portal.Azure.com to manage your services.
• 🔑 Create a new workspace by clicking the 'New' button and searching for 'Machine Learning Workspace by Microsoft'.
• 📁 Organize your Azure assets using resource groups, which are logical containers for billing and automation purposes.
• 📍 Choose a data center location based on where you need to consume the final output of your data.
• 💾 Azure Blob Storage is used for cloud storage, and you'll be charged based on the amount of data stored.
• 🔍 The workspace name must be globally unique and only contain lowercase letters to be used in the URL for cloud storage.
• 📈 For most users, the standard pricing tier is sufficient for initial prototyping and experimentation.
• 📊 The Azure ML workspace is a collaborative tool where you can invite team members and manage access to data, experiments, and models.
• 🔗 Notebooks are a new feature in Azure ML, allowing for coding in a web-based environment and supporting Python and R languages.

Q & A

What is the purpose of the video?

The video is an introduction to Azure Machine Learning, guiding viewers through the process of creating an Azure subscription and an Azure ML workspace, and exploring the features within that workspace.

How can one start a free trial with Azure?

To start a free trial with Azure, one can either use the provided link or search for 'Azure free trial' in a search engine. After clicking 'Start free trial', you will need to log in with a Microsoft account and provide a phone number for verification, along with a credit card for identity verification.

What is the significance of the Azure ML workspace?

An Azure ML workspace is a centralized location for managing and collaborating on data science projects. It allows users to create and manage experiments, access datasets, train and save models, and share these resources with team members.

What is a resource group in Azure?

A resource group in Azure is a logical container that holds related Azure resources for ease of management, organization, and automation. It's like a folder for your cloud-based assets, and deleting a resource group will delete all the resources within it.

Why is location important when selecting a data center in Azure?

The location of the data center is important because it can affect the cost and speed of data transfer. There is usually no significant charge for data ingress (into the cloud), but data egress (out of the cloud) can be costly. It's recommended to select a data center close to where the final output of the data will be consumed.

How is the Azure Blob Storage used within the context of an Azure ML workspace?

Azure Blob Storage is used as cloud storage for backing up data and saving Azure ML experiments. It incurs a cost of about $0.02 per gigabyte per month. If the storage container is deleted, the workspace won't be deleted but will be in an error-locked state due to missing data references.

What is the global uniqueness requirement for the name of an Azure ML workspace?

The name of an Azure ML workspace must be globally unique because it will be part of the URL for the cloud storage. It should only contain lowercase letters, with no symbols, numbers, or spaces.

What is the role of a web service plan in Azure?

A web service plan in Azure determines the level of service for deployed web services. It defines how robust the service should be, the number of people and transactions it should support, and the associated costs.

How can multiple workspaces be managed in Azure ML?

Multiple workspaces can be managed by switching between them within the Azure ML Studio interface. Users can create different workspaces for different projects or teams and invite specific members to collaborate within each workspace.

What is the Notebook feature in Azure ML, and how does it benefit data scientists?

The Notebook feature in Azure ML allows data scientists to write and run code in a web-based environment that supports multiple programming languages, primarily Python and R. Notebooks can be shared, deployed as web services, and are a popular tool for collaborative data science projects.

What is the process for inviting users to collaborate in an Azure ML workspace?

To invite users to an Azure ML workspace, the workspace owner can go to the 'Settings' section, select 'Invite users', and then enter the email addresses of the individuals they wish to invite. The invitees will be able to access the experiments, data models, and other assets within the workspace.

Outlines

00:00

🌐 Setting Up Azure ML Workspace

This video segment guides viewers through the initial steps of setting up an Azure ML workspace. The presenter starts by advising viewers to secure an Azure free trial subscription, either through a direct link or via a search engine. Once the free trial is accessed, the user is required to log in with a Microsoft-type email and provide a phone number and credit card for identity verification. The segment emphasizes that the credit card will not be charged beyond identity verification, though users in some regions may face a nominal fee. The presenter then proceeds to log into the Azure portal, highlighting the interface and various services offered, focusing particularly on Azure Machine Learning Studio and the steps to create a new workspace.

05:02

🔧 Configuring Azure ML Workspace

The second part of the video script explains how to configure the newly created Azure ML workspace. The presenter details the process of creating a resource group, choosing a data center based on data input/output needs, and setting up a new storage account essential for Azure Blob Storage. They discuss the importance of a globally unique name for the workspace, which doubles as a URL. The video also covers the creation of a new web service plan, emphasizing choosing the appropriate service tier based on expected transaction volume. Finally, the presenter describes how to pin the workspace to the dashboard for easy access and how to initiate and navigate the Azure ML Learning Studio through a preferred URL.

10:05

👥 Managing Azure ML Workspace

The final segment of the video script focuses on managing the Azure ML workspace. It introduces the capability to manage multiple workspaces, select different regions, and invite team members for collaborative projects. The presenter emphasizes the workspace as a collaborative tool similar to Google Docs for data scientists. The video also introduces Azure ML's support for Python and R Notebooks, and how these notebooks can be shared or exposed as web services. Key features like experiments, web services, and projects are discussed, explaining how to pin assets to projects and manage various team projects simultaneously. The video concludes with a prompt to watch the next video for a demonstration on creating experiments and managing data within Azure ML.

Keywords

💡Azure Machine Learning Studio

Azure Machine Learning Studio is a collaborative, drag-and-drop workspace where data scientists can build, test, and deploy predictive analytics models. In the video, it is the primary tool used for creating and managing machine learning experiments and workspaces within an Azure subscription. It is highlighted as a specific service to be utilized for data mining tasks.

💡Azure subscription

An Azure subscription is a service agreement that provides access to Microsoft Azure cloud services, including computing, data storage, and networking. In the video, the creation of an Azure subscription is the first step to start using Azure Machine Learning Studio. It is essential as it allows users to create and manage resources within the Azure ecosystem.

💡Azure ML workspace

An Azure ML workspace is a centralized place within the Azure Machine Learning Studio where data scientists can collaborate, store data, access models, and manage projects. The video demonstrates how to create such a workspace within an Azure subscription, which is a crucial step for organizing and conducting machine learning experiments.

💡Resource Group

A Resource Group in Azure is a logical container that holds related Azure resources, simplifying management, deployment, and access control. The video explains that resources within a group can be managed as a whole, which is useful for billing and automation purposes. It is used to organize the workspace within Azure.

💡Azure Blob Storage

Azure Blob Storage is a service for storing unstructured data, such as text and images, in the cloud. It is mentioned in the video as the type of storage account that will be used to back up data and save Azure ML experiments. It is an important component for data persistence and retrieval within Azure projects.

💡Data Center Location

The Data Center Location refers to the geographical location where the data center is hosted. In the context of the video, selecting the data center location is important because it can affect data transfer costs and latency. The speaker recommends choosing a location based on where the final output of the data will be consumed.

💡Storage Account

A Storage Account in Azure is used to store data that can be accessed by cloud services. The video specifies that a new storage account is created for the workspace, which will be used to store data and save machine learning experiments. It is a fundamental part of the infrastructure for Azure ML workspaces.

💡Globally Unique Name

A Globally Unique Name is a naming requirement for certain Azure resources, ensuring that the name does not conflict with any other resources within Azure. The video emphasizes that the workspace name must be globally unique, as it will be used in the URL for cloud storage access.

💡Pricing Tier

The Pricing Tier refers to the level of service and the corresponding cost associated with using a cloud service. In the video, the presenter suggests leaving the pricing tier as 'standard' for most users, especially during the prototyping phase of a project.

💡Web Service Plan

A Web Service Plan in Azure defines the compute resources that will be used when deploying a web service. The video indicates that for most users, the default or free tier is sufficient, especially when they are just starting out and their needs are limited.

💡Collaboration

Collaboration in the context of the video refers to the ability for multiple users to work together within the same Azure ML workspace. The workspace is described as a self-contained environment that can be shared with team members, allowing for collective work on data science projects, which is likened to a Google doc for data scientists.

Highlights

The video guides viewers on creating an Azure subscription and an Azure ML workspace.

Explores the features within the Azure ML workspace.

Provides a direct link to start an Azure free trial subscription.

Details the process of signing up for Azure, including the need for a phone number and credit card for verification.

Explains the Azure dashboard and how to navigate to Azure Machine Learning Studio.

Demonstrates creating a new workspace in Azure and the importance of naming and resource group selection.

Discusses the concept of a resource group as a logical container for cloud assets.

Advises on selecting a data center location based on data consumption needs.

Introduces Azure Blob Storage as a cloud storage solution for backing up data and saving ML experiments.

Highlights the need for a globally unique name for the workspace due to its use in a URL.

Mentions the option to leave the pricing tier as standard for most users.

Explains the web service plan and its role in deploying web services with specific tiers of service.

Provides a tip for directly accessing Azure ML Studio through Studio.AzureML.net for convenience.

Demonstrates how to invite users to a workspace for collaborative data science projects.

Notes the cost implication of $9.99 per workspace in Azure.

Introduces the concept of Notebooks in Azure ML for coding and sharing programming environments.

Details the Experiments tab as the central location for data science files in Azure ML.

Shows how to create and manage multiple projects within a workspace.

Encourages viewers to watch the next video for a walkthrough on creating the first experiment, importing, and exporting data.