Overview of tools for citizen data scientists in Azure
In this introductory chapter we will start by illustrating what Machine Learning can do for a business, and how the cloud
can be an ideal solution for Machine Learning. After that, we will shortly go over the different tools that are available
for citizen data scientists to do Machine Learning in Microsoft Azure.
- What is Machine Learning
- Why Machine Learning in the cloud?
- Machine Learning in Microsoft Azure
Introduction to Machine Learning
This classroom training does not require people to be familiar with Machine Learning. This introductory module makes sure
all participants have a common ground for diving into the rest of the training by discussing the basic concepts of Machine
- Which questions can Machine Learning answer?
- Machine Learning methodology
- Data preparation
- Classes of Machine Learning algorithms
- Model evaluation
Business Intelligence for many years focused on turning data stored in structured, relational databases into insights or actionable information.
There is however plenty of useful data that less easy to access such as plain text, images, phone recordings, ... . Cognitive services provides web services
hosted in Microsoft Azure to convert these sources into an easier to analyze format (mostly json documents). In this chapter we will give an overview of the different
cognitive services, where we will introduce the vision, speech, language, web search, and decision APIs. Some of these services are ready-made, whearas others are customizable.
- Overview of cognitive services
- Ready-made services
- Customizable services
Azure Machine Learning Service: Automated ML
Azure Machine Learning Service is a service that helps to bring Machine Learning to the enterprise level, for example
by offering tools that help with documentation, deployment, high availability and performance. This service contains tools
for data scientists, as well as data citizens. One of the tools that may be especially useful for citizen data scientists is
Automated ML, where Machine Learning is done in an automated way, with little time investment, programming skills or
domain knowledge needed.
- Introduction to Azure Machine Learning Service
- What is Automated Machine Learning?
- Configuring an Automated ML run
- Deploying and consuming an Automated ML model
Azure Machine Learning Service: Designer
A second service available in Azure Machine Learning Service is the Designer. This allows you to visually connect modules
to create Machine Learning pipelines using a drag-n-drop approach. A module is an algorithm that you can perform on your data,
such as a data transformation, training an algorithm, scoring new data, and validating a model.
- What is the Designer?
- Loading data
- Preprocessing data
- Creating Machine Learning Models
- Deploying models
AI features in Power BI
Power BI is a very popular tool for visualizing data. Lately, more and more features have been added, that allow for some
more advanced data analysis. Amongst others the Cognitive services and machine learning models created in the cloud can be consumed in Power BI Data Flows and Power Query.
- Introduction to Power BI
- Using ML models in Power BI Data Flows
- More machine learning options in Power BI
In this two-day course we will introduce the basic concepts of Machine Learning for citizen data science. We will walk through
a number of tools that can be used to create and deploy ML models in Microsoft Azure without a lot of Machine Learning or
coding knowledge. We will see Azure Machine Learning Service, in which you can either let your models
be created automatically (Automated Machine Learning), or where you can create your ML pipelines using a
drag-n-drop approach (Designer). We will have a look at different Cognitive Services, which are AI services and cognitive
APIs that you can easily use to built intelligent apps. Finally, we will see how to consume these models in Power BI.
This course is intended for people who plan on using more advanced data analysis techniques. This can be BI developers as
well as data analysts. Also project managers who which to get a better overview of Machine Learning possibilities in Azure
can benefit from this course. Students should have a general background in working with data, and some experience
with business intelligence in general.