Focus of this one-day workshop will be building data-driven
analytical solution for enterprise environment. Using Microsoft on-prem
and in the cloud solutions, the workshop will guide you through building
a solution in either of environments, as well as a hybrid solution,
getting best of both worlds. Building working end-to-end solution with
given technology will be main focus of the workshop and not so focusing
on exploring different machine learning algorithms.
Building
on-prem solutions with Python (and some R) starting with overview on
what is available in SQL Server 2017, continuing with 2019 features and
big data clusters and exploring the capabilities of data virtualization
and the power of Polybase language.
Moving to cloud services, we
will be exploring the automated machine learning services, Azure
Databricks, Azure Data Factory, Azure SQL Datawarehouse and Cosmos DB
with stream analytics.
Last part will be focusing on ML.NET, as part of creating and building .NET applications using Machine Learning.
Workshop
will be using mainly one (in addition another data-set for streaming
analytics) and covering different machine learning scenarios based on
better services (on-prem or in cloud).
Program of different modules
Module 1: 9.00 -
10.00 Using Python and R on-prem. Module will introduce Python and R
with SQL Server 2017/2019 in-database integration, using native scoring,
building and deploying the models. (Module: 30% presentation; 70%
Hands-on)
(10 min break)
Module 2: 10:10 – 11.00 SQL Server
2019 Big Data Clusters Module will show and use the new functionalities
of Big Data Clusters in SQL Server 2019, using Azure data studio and
bringing together data from different sources, using Polybase and
Jupyter notebooks in order to build the enterprise analytical solution.
(Module: 50% presentation; 50% Hands-on)
(10 min coffee break)
Module
3: 11.10 – 12.00 - Machine Learning services with Python; Module will
show you the capabilities of using automated machine learning project as
part of a service, creating pipelines and deploying the model. (Module:
30% presentation; 70% hands-on)
Lunch: (30 min Lunch break)
Module
4: 12.30 – 13.30 Using Azure Databricks for data analysis – Module
will walk you through building a complete machine learning environment
for collaborative and shared projects in Databricks. Using Python,
Scala, R or SQL, we will be building, deploying and consuming a
predictive model.
(Module: 30% presentation; 70% Demo/hands-on).
(10 min break)
Module
5: 13.40 – 14.50 Azure Cosmos DB and stream analytics – Module will
look into using and setting up Comsos DB, Event hub, Azure functions and
Stream analytics for viewing and running analytics on stream data. With
simple data analytics and simple model predictions, you will be able to
predict on stream data in the cloud.
(Module: 50% presentation; 50% Demo/hands-on)
(10 min break)
Module
6: 15.00 – 16.10 Azure Data Factory, Azure SQL Datawarehouse – Module
will explore the data wrangling capabilities with ADF, storing the cold
path data in Azure SQL Datawarehouse and some help of Databricks and
using Power BI to do end-to-end data analysis.
(Module: 40% Demo, 60% Demo)
(10 min break)
Module
6: 16.20 – 17.00 Simplified ML.NET – Module will walk you through
simplified Machine learning for .NET developers, helping developers
using ML without leaving the .NET ecosystem. At the same time bringing
.NET solutions with ML solution integrated.
(Module: 60% Presentation, 40% Demo)
Target Audience:
This
workshop is intended for Data Analyst, data scientist and data
engineers that are building ML projects on-prem but would like to extend
their process into the cloud and extend the capabilities of cloud
computations. Invited are also business people, domain experts, who
would like to get a better knowledge on new approaches in Azure cloud
and SQL Server 2019.
Outline:
This one-day
pre-con will be demo based and users will be able to tag along. It is
encouraging to bring your own laptop and have free azure subscription
activated. You will need a laptop running SQL Server 2019 (Developer
edition) and Azure Data Studio and Visual Studio with SSTD, Azure
Storage explorer and Microsoft SQL Server Integration Service Feature
Pack for Azure. Script and code will be available upon session on
Github; same datasets will be used on all scenarios and will be
distributed as the beginning.
Take aways:
Participants
will walk away with understanding how Analytics in Azure can be done,
what are the benefits of SQL Server 2019, basic knowledge on Python
Language, and all the relevant processes that enterprise analytics need:
data manipulation, data preparation, predictive analytics, model
deployments and consumption.