Applied Data Science in Azure with Microsoft Fabric and Azure Machine Learning

Applied Data Science in Azure is a full-day training that will deliver a look into the world of data scientist that rely on using Microsoft Fabric and Azure Machine Learning on a daily basis.

This training will give attendees insights on a powerful Azure duo: Microsoft Fabric and Azure machine learning. We will learn where and how to start with each service, how to build data pipelines and data flows, perform data analysis tasks and of course build, deploy and consume predictive models. At the end of this training attendees will have sufficient knowledge to use both Microsoft Fabric and Azure Machine Learning to make better decisions - and train machine leaning models – for their organizations.

The full-day training consists of 6 modules and will explore all the important assets, tools and best practices for machine learning workloads in both Microsoft Fabric and Azure Machine Learning.

Short description
Applied data science in Azure with Microsoft Fabric and Azure Machine Learning is a full-day training, that will deliver a look into the world of data science using both technologies for machine learning tasks. From building and deploying predictive models with dedicated built-in tools, to using the Python SDK and other open-source frameworks.

What you will learn at this full-day training

• Understand the basic concepts of data science processes and what cloud services bring to the table
• Get on board with designing and preparing Machine Learning solutions by using both Azure Machine Learning and Microsoft Fabric
• Get familiar with notebooks using Python and using Spark
• Learn how to explore data and train and tune predictive machine learning models
• Learn how to prepare your models for deployment by creating pipelines and using MLFlow
• Deploy, monitor and retrain your models
• Learn how to use the Python SDK and Spark
• Learn how to use the pre-build and pre-trained models and datasets available in Azure Machine Learning

Training modules
The time outline for the training is designed in 6 modules each for 70 minutes.
Each of the modules will focus on tasks in both Microsoft Fabric and Azure Machine Learning.
Times are displayed in UTC. Coffee and lunch breaks will be aligned with the organisers on the day of the workshop.

08.45: Gathering and preps
09.00 - 10.10: Module 1
10.10 - 10.15: Short Break
10.15 - 11.25: Module 2
11.25 - 11.35: Coffee Break
11.35 - 12.45: Module 3
12.50 - 13.50: Lunch break
13.50 - 15.00: Module 4
15.00 - 15.10: Coffee Break
15.10 - 16.20: Module 5
16.20 - 17.30: Module 6

Modules:

Module 1
Starting with light general introduction to Microsoft Fabric and Azure Machine Learning Services (AML), and how you navigate these services. We will also cover the basic key concepts and how they affect both offerings and subscriptions. We will cover the key differences and similarities between both services and explore additional administrative settings.

Module 2
This module will be dedicated to data storage and compute offerings. For Microsoft Fabric we will cover the outstanding OneLake data store, learn how it works, what is is capable of and how to get navigate. In Azure Machine Learning we will look into different Azure data store options like ADSL Gen2, and learn about creating datastores and datasets.

Module 3
Data Movement, data pipelines and data transformation will be key focus in this module. For both offerings we will look and explore the usage of notebooks with Python or Spark. In addition, for Microsoft Fabric, we will look into the UI-based and code-based data engineering tasks and explore the use of shortcuts and data lakes. With Azure Machine Learning we will look the designer and learn how to build pipelines using the Python SDK.

Module 4
Focusing on data analysis using notebooks for both Microsoft Fabric and Azure Machine Learning will be first part of this module. In the second part, we will build prepare and build a machine learning model and run experiments. We will also use Power BI for exploratory data analysis in Microsoft Fabric, and use notebooks for Azure Machine Learning

Module 5
Once we have the machine learning model created, we will test it’s predictions, and perform inference. In both Microsoft Fabric and Azure Machine Learning will schedule and run jobs using MLFlow to retrain the model and expose it’s endpoints for consumption.

Module 6
In the last module we will be wrapping up both services by creating an end-to-end integrated solution. On top of this we will explore various options of visualizing the results of our analysis and model predictions.

Key takeaways
Learn how to use Microsoft Fabric and Azure Machine learning services for machine learning tasks and be able to build an end-to-end machine learning solution from a scratch.

Target Audience
Data Scientists, Statisticians, Machine Learning Engineers

Broader Audience
Data Analysts, BI Analysts, Big Data analysts, Data engineers, Data architects, Tech Leader, DevOps Engineers, and Business Leader.

Prerequisite knowledge for attendees
Some background in Machine learning or statistics. Any additional knowledge of Azure Machine Learning or exploratory data analysis is an added benefit while attending the workshop.

Technical prerequisite for attendees
• Working laptop with admin access (Win or Mac)
• Installed Visual Studio Code and Python Environment
• Conda environment and additional Python packages installed (Pytorch, ONNX, ...)
• Access to the internet
• Credentials and credit (free credit) for accessing the Azure portal
• Material and demos

All materials (Markdown, iPynb, Bicep, Pytorch, Py) and accompanying materials will be handed to attendees before the workshop. Material is prepared for self-paced learning.