The Truth About Application Release and Deployment - Top 10 Myths Exposed
Predictive Analytics and Azure Machine Learning Case Studies
1. MISSION
Pragmatic Works was sought out to assist a large Insurance provider
in Florida to identify trends in their data to make better business
decisions with several lines of business, including Rental Policies and
Extended Warranty Policies.
RESULT
Pragmatic Works facilitated the use of Azure Machine Learning and
Azure Data Lake to analyze the policy data and developed models
Insurance Company - Miami, FL
CASE STUDY
Pragmatic Works Solutions with Predictive
Analytics and Azure Machine Learning
Technology
• Azure Machine Learning
• Hive
• Azure Data Lake
which provided greater insight to the correlation between different data elements and customer groups.
The machine learning experiments provided previously unknown information regarding the relationships
between various groups of consumers and their policy history and use of insurance products. New insight was
gained regarding the correlations of consumers and products which drive business practices. Azure Machine
Learning helped identify relationship variables used to determine sales information.
Using the data provided in the experiment, allowed the company to better direct their resources from methods
which would not provide an increase in sales to areas where the resources would provide more impact. This
modification redirected the insurance company’s planned investment from an area which would have yielded
minimal ROI, to an area where the resources could be put to more effective use.
Academic Testing Company - Princeton, NJ
MISSION
Pragmatic Works was contracted bya large academic testing company
that operates centers for standardized testing to use machine learning
to identify patterns in their data to help expose fraud and patterns of
cheating.
RESULT
Using machine learning to access data stored in Azure Databases,
various experiments were developed to uncover patterns of possible
fraudulent activities and areas of susceptibility. One of the criteria
which was very important to the client was to be able to run the
experiments with large amounts of data. To accommodate this
Technology
• Azure Machine Learning
• Power BI
• Azure SQL Database
• Azure Event Hubs
• Azure Data Factory
request, several experiments were deployed as web services so that they could be run in Data Factory. The
results of the experiments were deployed to Azure Event Hubs and run with Azure Data Factory. A graphical
representation was available via Power BI.
2. sales@pragmaticworks.com | pragmaticworks.com
Insurance Company - Manchester, NH
MISSION
PragmaticWorks helped the analytic and actuarial department oflarge
insurance company assess what kind of an environment would allow
them to extend their analysis and predictive analytics capabilities.
They wanted an environment which would allow them to grow and
expand their current development. The development is primarily in
Open Source R, and they were looking to determine what factors
and features could be provided to help them extend the capabilities
of their code to an environment where the code could be deployed
using more standard development methodologies.
RESULT
Technology
• Azure Cloud
• Microsoft “R”
• Azure Machine Learning
• SQL Server 2016
• Azure Data Factory
After Pragmatic Works familiarized the insurance company with the feature sets available in SQL Server
2016’s R Server and the capabilities of the Azure Cloud implementation of virtual machines, the insurance
company decided to make a large investment in the Azure Cloud for their analytics and actuarial department.
The capabilities of SQL Server 2016 to run R code and Power BI for exposing the visualizations created in R,
convinced the insurance company to migrate from their local open source solution to Microsoft R Open. This
decision resulted in a change to where some of the data will be stored as it will be migrated from Oracle and
Teradata to SQL Server 2016. The capability of Azure Machine Learning to deploy the R code using Azure
Data Factory to call a web service created by Azure Machine Learning was also explored.
The insurance company was most impressed with the capability of R to use not only server memory but
the ability to use disk, providing the capability to analyze much larger datasets, providing a significant
performance improvement. The solutions were deployed by a Pragmatic Works Microsoft MVP into Azure
Cloud. Pragmatic Works also introduced Azure Machine Learning as a platform to enhance insight derived
from data.
CASE STUDY
To learn how Pragmatic Works can help your company leverage Predictive Analytics
and Azure Machine Learning, please contact sales@pragmaticworks.com