Chapter 2. ML Studio Inside Out

While working on a predictive analysis model, you typically follow different steps, such as pulling data from one or more sources, exploring and preparing data, or applying different algorithms to get your desired output. Then, you test and improve on it. Usually, this is an iterative process. Once you are happy with your model, you find ways so that it can be deployed for production and other people or applications can consume or make use of your developed model.

To perform the preceding tasks, you need an environment with the right tools available. ML Studio provides you with everything to develop and deploy a predictive model.

In this chapter, you will start exploring ML Studio after you know how to create a Microsoft account and a Azure ML workspace. Then, you will get introduced to different parts of ML Studio and learn how to create an experiment. You can also find out, briefly, how to work with other projects in ML Studio collaboratively. This chapter...

Introduction to ML Studio

ML Studio gives you an interactive visual workspace to easily build, test, and iterate a predictive analysis model.

You drag-and-drop datasets and analysis modules onto an interactive canvas, connecting them together to form an experiment, which you submit to ML Studio to run or execute. To iterate your model design, you edit the experiment, save a copy if desired, and submit it again.

There is no programming required for this; visually connecting datasets and modules to construct your predictive analysis model is enough. However, if you need more functionality than what is available visually in ML Studio out of the box, you can write R or Python code to get the desired result. R or Python programming is not an absolute must to work with ML Studio.

https://azure.microsoft.com/en-gb/documentation/articles/machine-learning-what-is-ml-studio/

Before you start working with ML Studio, you need to get a subscription for Microsoft Azure and sign in to ML Studio. The following...

Getting started with Microsoft Azure

Getting into the details of Microsoft Azure is beyond the scope of this book. However, the following subsection details the steps to start with it by creating an account and starting a subscription.

Microsoft account and subscription

If you don't already have a Microsoft account, you need to create one by visiting http://www.microsoft.com/account. This URL might change in future and if so, you can just search online for Microsoft Account to find the right URL.

At the time of writing this book, if you sign up for the first time, Microsoft offers you a free trial for a month and a credit worth $200 to spend on the services on Azure, which is more than enough if you just need to follow through the examples in this book and use only ML Studio.

Once you are successfully signed in, you can visit https://manage.windowsazure.com/ to find different services available through Azure.

Creating and managing ML workspaces

You can scroll through on the left-hand side of the...

Inside ML Studio

You usually land at the ML Studio home page that contains a bunch of links to different resources, including documentation and quick-start videos.

Apart from ML Studio Home, you will also find the following tabs on the left-hand side of the screen:

• EXPERIMENTS: These are the experiments that have been created, run, and saved

• WEB SERVICES: This is a list of experiments that you have published

• DATASETS: This is a collection of all the datasets that are either uploaded or saved from a experiment along with all the sample ones

• TRAINED MODELS: This is a list of all the trained models

• SETTINGS: This is a collection of settings that you can use to configure your account and resources

Experiments

You can think of an experiment as any analysis you would perform in ML Studio—it can be a simple one, such as a simple statistical analysis, or a complex predictive analysis. An experiment inside ML Studio is a collection of modules connected hierarchically. A module is a unit that encapsulates...

Workspace as a collaborative environment

Workspaces enable groups to work on common projects by gathering data, modules, and experiments together in a single location for common use. Workspaces let users securely share ideas and resources. You can be a member of several workspaces and can easily switch between them.

As the owner of a workspace, you can invite others to the workspace by clicking on the Setting icon on the left-hand side of the screen and then clicking on USERS from the top tabs. You can invite others to the workspace by adding their Microsoft accounts.

Once you have successfully added other users, they can use the same workspace like you as an owner can, except that they can't invite others unless you give them ownership privileges.

One user can be an owner or user of more than one workspace.

Note

Note that a workspace can be shared and owned by multiple users, but billing is made only to the user who created the workspace.

Summary

Practically speaking, ML Studio is the Microsoft Azure Machine Learning! If you are working on a predictive analysis, ML Studio provides a platform for everything—for development, testing and deployment. It does this in the easiest way possible, just by mouse clicks.

In this chapter, you started with creating a Microsoft account and creating an ML workspace. Then, you explored ML studio from inside out. You moved on to create a simple experiment in ML Studio and also quickly explored how ML Studio can be used as a collaboration environment to work with others.

Now that you know ML Studio, in the next chapter you will learn about data exploration and data visualization using ML Studio.