Innovation-Day: Machine Learning - Hands on Lab

This content is designed for audience without any prior Machine learning knowledge. It starts from very basics and goes to advanced topics. We will try to keep this content live and include more and more advanced lab sessions with real life scenarious. Thanks for your support and feedback to make this content better.

Objectives

After completing this lab, you'll be able to:

• Understand the fundamentals of Machine Learning, statistics and data analytics principles
• Learn how to apply these concepts on a real-world scenario leveraging the power of Azure Machine Learning Studio from scratch
• Consume the results of our Machine Learning experiments in innovative ways via Power BI visualizations and conversational bots

Prerequisites

• Own any Microsoft Azure subscription
• Basic knowledge of SQL queries

Overview of the lab

The purpose of the lab is to guide you through the creation of a very simple Machine Learning prediction experiment using Azure ML. The scenario would be the following: you are getting ready for the next season of your NBA Fantasy League, and you want to predict the Player Efficiency Rating of every player, so you have some insights to pick the best players for your team.

This lab contains 7 chapters, the first six will cover the whole Machine Learning experiment creation and are the mandatory part, and the last one is a bonus track so you can understand different ways to interact with the results of the experiment after its completion.

Begin Hands on Lab

Detailed contents of the HOL

1. Chapter 1 – Getting things ready: Sign-up/sign-in to Azure and AzureML Studio

2. Chapter 2 - Create project and import data

3. Chapter 3 - Data prep & cleanse

4. Chapter 4 - Train and evaluate your model

5. Chapter 5 - Refine the model and re-evaluate

6. Chapter 6 - Predicting the future

7. BONUS TRACK - Visualize data with Power BI and also interact with it from a conversational bot

Closing remarks

Wrapping it up, there are some points to recap at the end of the lab:

• The 80/20 rule: data preparation and cleaning is the most important thing on any Machine Learning experiment, and it takes most of the time of the experiment creation
• Trying different approaches and adjustments use to improve the performance of the model, but it's a try-fail learning process…keep trying as long as necessary!
• Business/domain savvy is crucial to achieve the experiment goals, as the technical skills might not be enough to identify the most relevant features or the filters necessary to improve the quality of the dataset

References

• Azure Machine Learning Studio tool (http://studio.azureml.net/) and documentation ( https://docs.microsoft.com/en-us/azure/machine-learning/)
• Microsoft Professional Program for Data Science – https://academy.microsoft.com/en-us/professional-program/data-science/
• Basketball reference (source of the datasets) - http://www.basketball-reference.com/leagues/NBA_2017_advanced.html