Building Industry-Aware Analytics Solutions Using
Building Industry-Aware Analytics Solutions Using
Building Industry-Aware
Analytics Solutions Using
Azure Synapse Analytics
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 2
The primary audiences for this paper include architects and technical decision-makers mandated
to build industry-aware analytics solutions for their organizations. This may include chief data
officers, line-of-business application owners, data scientists, and data engineers. Technically oriented
business decision-makers will also benefit from this guide.
Through this guide, you will gain a high-level understanding of Azure Synapse, especially as it relates
to building industry-aware analytics solutions relevant to your industry. While the retail industry is
used to illustrate the concepts, the approach is applicable to any other industry as well.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 3
Rajeev Jain is a Senior Product Marketing Manager with Microsoft Azure Data & AI. In his current
role, he drives the business growth of the Azure analytics business through the industry and partner
lens. In the past, Rajeev has led sales, marketing, and partner businesses at Microsoft, and has
worked in the oil and gas and financial services consulting industry outside of Microsoft.
Editorial credits
Azure Data Engineering: Matt Burdick, Santosh Balasubramanian, Gautam Bharti,
Aaron Merrill, Trevor Olson
Building Industry-Aware
Analytics Solutions Using
Azure Synapse Analytics
5/ 9/ 15 /
Data and analytics are Understanding database Implementation
at the heart of digital templates in Azure Synapse
transformation
6/ 11 / 21 /
Overview: Industry–aware Industry scenarios enabled Summary and conclusion
analytics with Azure Synapse by Azure Synapse:
Retail example
8/ 13 /
What makes Azure Synapse Solution architecture
industry–aware?
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 5
Consider the retail industry. This industry was already adopting e-commerce rapidly due to the
cascading impact of digitization on their supply chains, operations, and customer experiences. The
pandemic led to the mainstreaming of many niche trends like contactless shopping and curbside
pickup experiences. The increased digitization of processes and experiences led to ever greater data
volumes from applications and interactions. This could potentially be combined and analyzed in
interesting new ways to generate actionable insights, often in real time, for driving business agility.
For example, the ability to capture data around customers’ online purchases could be combined with
their demographic profile and social media interactions to proactively generate personalized offers
or reduce churn.
Similarly, analytics can be applied to all areas of the business to convert data into insights for agile
decision-making and execution. In practice, rapidly creating analytics-based applications, like ones
that enable personalization, demand forecasting, or supply chain optimization, needs a robust data
foundation. This foundation is often a challenge to create due to the multiple internal and external
systems not talking to each other, thereby creating data silos. Unlocking and combining data
trapped in these systems is hard. A common data blueprint is needed to avoid each scenario being
an independent data project that slows down digital transformation and reduces innovation.
Azure Synapse provides the foundational analytics and data shaping capabilities for your business—
so that data from different applications and services across different business areas (like customer,
product, sales, marketing, and operations) can leverage the same connected dataset and generate a
holistic view.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 6
Azure Synapse provides several different components, enabling customers with an integrated and
seamless experience to build and operate their data projects in a secure and scalable way:
• An analytics “hub” that provides the limitless capability to analyze any type of data
independently of the underlying storage technology. Customers can take advantage of the
industry-leading SQL, big data, or data discovery analytics services that—combined with
cost‑efficient storage capabilities—add semantics and context so that data can become
an organized strategic asset.
• The comprehensive selection of prebuilt database templates for customers in different industries.
These templates enable you to start your data projects with application-independent data
structures. Leveraging proven database templates built with decades of experience across
industries reduces project risk and helps accelerate time to value.
• Synapse pipelines give the customer an integrated low-code/no-code experience to move and
transform data from silos into a unified structure with hundreds of prebuilt connectors at no
additional cost.
• SQL and Spark analytics compute engines for data warehousing and big data analytics.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 7
Find out more about the specific services at Azure Synapse Analytics:
With the availability of database templates within Azure Synapse, you can shape and describe data
in a standardized way, for your specific industry. Database templates also make the integration
of different data sources easier. New and existing projects can leverage out-of-the-box database
templates that address a wide variety of industry-specific business areas. Finally, the database
templates have been built with an ecosystem in mind. You can rapidly build analytics-infused
industry use cases by customizing and extending the standard templates using the database
editor in Azure Synapse:
With the database design capabilities in Azure Synapse, you can extend and customize database
templates to fit your specific scenario. When the database templates are deployed, the metadata
(including descriptions, data types, and relationships) is directly attached to the data lake. Once data
across heterogeneous sources is ingested and structured in a standardized way, it can be exposed
to seamlessly interact with the analytics engines in Azure Synapse (Synapse SQL and Spark), as well
as with Power BI. This enables the rapid creation of analytics applications and makes it easy for data
analysts, business users, and AI/ML experts to do their job.
Database template-aware components in Azure Synapse, such as pipelines or SQL on-demand, can
be used to automate the data lifecycle, including ingestion, exploration, and analysis. This increases
productivity and reduces time to insight.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 11
In retail, offline merchants are embracing new business models, like contactless shopping and
curbside pickup, while e-commerce has been accelerating for several years already. Being able to track
and analyze increases in the number of interactions due to e-commerce—before, during, and after
purchase—is becoming essential for retailers. Customers not only expect a friction-free experience
but also appreciate relevant information sharing in an over-communicated world, versus random
messages and offers.
Figure 5 shows a schematic of a personalized offer generation and serving system. The system
provides a unified set of capabilities to integrate data from multiple sources (like the customer’s
prior interactions across physical and virtual channels), develop useful customer analytics, and
enable targeted offer generation via a campaign management system.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 12
Azure Synapse is at the heart of such a system, with the retail database templates providing an agile
way to build such a capability:
Figure 5: Personalized offer generation system with Azure Synapse database templates
The template being used in Figure 5 is just a small selection of customizable tables that are available
and needed for the personalized offer generation use case.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 13
Solution architecture
To build an industry-aware solution (like the retail example in Figure 5), there are several questions
that need to be answered before you can start with the implementation. The answer to those
questions also helps to outline the solution architecture that is used for the specific project:
• What is the end goal? In our example, we want to generate relevant offers so that customers
feel that we know them intimately. These offers should be served across the channels (for
example, in-store, online, or telesales) and be based on the buying and browsing history of the
customer.
• What data is needed? The next logical step is to identify the different datasets that you need
to build such a solution. Based on what needs to be achieved, there are different data sources
that need to be tapped and relevant data combined. In our case, data from the e-commerce site
(transactions and browsing history), the POS system, general customer information, and product
information need to be used to generate an appropriate solution.
• What tools are needed? Finally, we need to decide what tools should be used to bring the data
together, analyze data, and feed the relevant information or insights to different upstream and
downstream systems.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 14
Azure Synapse provides an integrated environment that provides data integration, data shaping, and
analytics engines to rapidly build an industry-relevant solution:
For our solution, Azure Synapse pipelines are used to connect to the source systems and a lake
database is used to bring all the information together in a single location. Database templates build
the foundation for the database where all the different data sources are combined.
A Spark notebook is used to analyze data and make recommendations for customer offers that are
stored in the lake database where different systems can pick it up. All this tooling is integrated within
Azure Synapse Studio, a unified environment to develop, test, and get the solution into production.
Thanks to deep integration with Power BI and the fact that data sits with all the relationships and the
associated metadata in the lake database, data analysts can themselves access, analyze, and derive insights.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 15
Implementation
In this section, we will demonstrate how convenient it is to build an industry-aware analytics solution
with the capabilities of Azure Synapse. We will start by creating a lake database.
In our case, we will use the retail template and start selecting the appropriate tables from the
business areas that are relevant to the desired retail industry solution (for example, a personalized
offer solution).
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 16
Different business areas, like customer, sales, finance, and marketing, can be found on the left panel.
These business areas have hundreds of prepopulated business entities (tables) that can be used as
basic building blocks to customize and create a lake database layout that can easily be extended in
the future:
As soon as you have selected the entities you are interested in to start with, the Create database
action will bring you back into the Azure Synapse workspace, where a new lake database is created
that can be fine-tuned to your specific need.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 17
Entities and attributes can be added, changed, or removed to reflect the changes in your business
processes, while the service will keep the lineage back to the original database template in case you
need to add information in the future or want to use a newer version as the basis for your model:
For each table, the Azure Synapse database designer allows you to change the name, attributes,
and relationships as shown in Figure 9. However, for the most impactful AI and ML integration, we
recommend minimizing changes and staying as close to the original database template as possible.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 18
Data processing pipelines provide connectors to different sources, which makes it easy to connect to
different source systems as shown in Figure 10. There are hundreds of different connectors available
that can be used out of the box to get data into the new lake database easily (for example, SAP ECC,
Dynamics CRM, Magento, and others):
Additional information about available connectors can be found at Connector overview – Azure
Data Factory & Azure Synapse | Microsoft Docs.
The pipelines also provide an easy way to join, filter, select, and modify data that is coming in, as you
can see in Figure 11. The graphical user interface makes it easy for everyone to understand what is
happening with the initial data. A data lake sink understands the format of the new lake database
using rich metadata from the template that has been used. No additional configuration is needed, as
the service already knows how and where to store data from the lake database that was created:
Once you have landed the data in the common format, you can integrate visualization or ML models
that work directly out of the box with the format of your data as shown in Figure 12. They can
leverage the data stored in the lake database and provide data back, as well. Leveraging database
templates as the blueprint for the lake database helps to integrate these components easily as the
solutions already know the format of the data.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 20
In the retail example, the ML solution will create personalized promotional offers and advertising
events that are stored back in the lake database where they can be picked up by other solutions:
Now that the data is available in a standardized format and enriched with metainformation, it can
be exposed and viewed by different services. Data processing pipelines can be used to publish the
newly generated insights into other services. The connectors mentioned earlier can be used for this
as well (for example, a web shop, the POS system, or social media channels, as depicted in Figure 5).
In addition, the data can also be exposed over a SQL interface provided by Azure Synapse out of the
box on top of the created lake database. SQL on-demand can be used on top of the lake database
and make use of the rich metadata stored in the model. This enables every solution that can connect
to SQL to connect and access the data in the lake.
The data expressed in the lake database in Synapse can be used in Power BI and other Power
Query‑enabled apps (for example, Power Apps and Customer Insights) as well as using the
Synapse Power Query connector. With the ease of data import, along with the automatic import
of relationships and the additional metadata in the lake database (names, descriptions, and other
metadata), business users can use Power Query’s no-code transformation capability to get faster and
better results for their ad hoc business needs. It is easier than ever for business users to select the
right data across tables, by leveraging relationships, and create queries with joins and filters without
needing to write code or understand relational database concepts.
Building Industry-Aware Analytics Solutions Using Azure Synapse Analytics 21
For more information and implementation guidance, check out the Azure Synapse database
templates documentation.
• Create an Azure account to create a new Azure Synapse instance and explore your data instantly.
• Visit the Azure Synapse Analytics documentation webpage to see how cloud analytics can help
your business.
• Request a call from an Azure Synapse Analytics sales specialist if you have any questions
or need help.