Azure Essentials: Data modeling and Business Intelligence

Azure Essentials: Data modeling and Business Intelligence

[Serious Music] – Welcome to Azure Essentials. In the next few minutes
we’ll explore the services Azure offers to help you
model and visualize your data for interactive reporting
and business intelligence. Wherever your data is stored, Azure offers Azure
comprehensive lineup of services to help you ingest, transform
and store the aggregate data so that it can be modeled and explored with commonly used visualization tools, like Microsoft Power BI, Excel and more. Azure’s an open platform.
We want it to give you all the tools that you
need to make sense of the broadest range of data sources, whether you’re working
with well structured operational databases hosted
on-premises or in the Cloud; online data sources, for example
data from your Cloud apps and other reporting services; stores of unstructured
data, such as event logs, emails, documents, media files and more; or even realtime data streams, including telemetry from
sensors and IoT devices. Let’s start with data ingestion. If you’re working across
disparate data sources, you’ll want to first
extract the data you want from each source, transforming
it so that it can be unified and aggregated with other data, and stored in a central data repository. Able to scale to hold all of the data that you might want to report on. If you’ve got structured data in operational databases on-premises, you can leave your data where it is, or you can move it to one of
Azure’s many database options. We’ve got a gallery of
data connectors available to connect to Azure, or
you can develop your own. If you’re working with large
amounts of unstructured data that you’d like to report on, Azure has the scalar compute
power and analytics engines to extract information and
transform it into a state suitable for storing and modeling. Now to ensure that your
reporting stays up to date, it makes sense to automate the ingestion and transformation process on a schedule that suits your needs. Azure offers a range of ways to do this, including scripting your own
batch processes in Azure CLI, or using Azure Data Factory; a full hybrid data integration service. Azure Data Factory allows you to create, schedule and orchestrate your ETL, or Extract Transform
Load workflows at scale, wherever your data lives; in the Cloud, or self hosted on-premises. Now collecting massive volumes of data from multiple structured
and unstructured sources and ingesting data from
all of them in parallel requires a powerful
storage service to give you the foundation for reporting
and business insights. In Azure, Azure SQL Data
Warehouse has the performance, elasticity and compute power
to scale up or down as needed, allowing you to host a
centrally curated repository virtually unlimited in size. But a full data warehouse
solution’s not always required. If your total storage
requirements aren’t gonna exceed four terra bites, Azure SQL server is likely to be sufficient for your needs. A Data Model is the heart
of any analytic system and establishes the structure of your data and the relationships between your data. Whatever the data source,
models provide data views that can be explored for
business intelligence; therefore needs to be
capable of scaling out to user demand and fast. In Azure we focus on enabling you to combine all relevant data into a rich semantic business
oriented data model. Able to cache potentially
many billions of rows of data in memory for lightning fast queries. With Azure you can create this
semantic data model directly in Power BI for an all in
one self service solution, or you can use Azure Analysis Services to build out an IT managed model, accessible by Power BI and
other data visualization tools. Modeling in Power BI or
Azure Analysis Services, is very similar. In each case you create a semantic model, that presents the data as
logically linked tables, mashed up from different sources, presenting a unified, consistent
view of the underlying data using friendly names,
understandable to the business users who’ll be querying this model. Using data analysis expressions or DAX, you can also derive
new data for the model. You can use this to filter the data or calculate and create
new measures and KPIs. One thing to note, there are times where rather than ingesting the
data into a data warehouse, you may ingest data
directly into the model for ad-hoc data modeling. This might be the case
when you’re working with unstructured datasets, where
you’re still exploring data for patterns and structures
not yet supported by the warehouse schema, or where your data needs to stay on-premises. Now if you’re interest in a hybrid model, or ultimately moving your data to Azure and you’ve worked with
Microsoft SQL Analysis Services, or SSAS on-premises, you’ll find working with these data models very familiar. Moving from an existing
on-premises SSAS implementation to Azure’s straightforward. You can move your
on-premises model to Azure and continue to link to your
existing on-premises databases via the Azure Gateway; or you can migrate your on-premises databases too. All you need to do is
give your service a name, specify what capacity you need
and the geographic location where you want it hosted;
and Azure handles the rest. Unlike on-premises implementations, it takes just a few minutes
to spin up a new instance of Azure Analysis Services
and provision more deep provision capacity, to
meet change in business needs. Now it’s in the Cloud, you no longer need to do the capacity planning in advance. You can scale up and out to meet demand. Additionally, with Web designer, you can now connect to the data warehouse and other sources to
develop your data models. With the data model in
place, you can now visualize your data using your visualization tools of like Power BI. Power BI’s a versatile and easy to use way for everyone in your
organization to access rich, interactive reports, and to gain insights from the data from many angles. Using Power BI dashboards
you can create single panes showing complete 360 degree views of your business, project, or department. And you can also provide
realtime streaming of data into these dashboards
from social media sources, service uses metrics or
sensors and IoT devices via Azure IoT services. These dashboards help
everyone in the organization make data driven decisions
and share their other insights with other people and teams. Each user can then build on
top of this as they wish. They don’t even have to be familiar with how to use Power BI. Power Bis support of
natural query language, which means users can
interrogate the model with a Q and A capability,
the data view they need, without having to create a report first. Users don’t even need to type,
as Azure Cognitive Service Integration into Power BI
includes speech recognition too. So that was a quick overview of modeling and visualizing your data
for instructive reporting and business intelligence in
action. Keep checking back on Azure Essentials for more topics, and check out our hands on learning series at the link show. Thanks for watching.

Comments (3)

  1. Thanks Matt – nice short overview of what can be a daunting, deep and broad range of services. Certainly going to help me with pruning elements of a presentation I'm prepping for next week.

  2. This was an extremely helpful video – thank you.

  3. Matt my man! You've dazzled me once more with the power of Microsoft tools

Comment here