###BeginCaseStudy###

Testlet 1
Background
Overview
Northwind Electric Cars is the premier provider of private, low-cost transportation in Denver. Northwind drivers
are company employees who work together as a team. The founding partners believe that by hiring their drivers
as employees, their drivers focus on providing a great customer experience. Northwind Electric Cars has a
reputation for offering fast, reliable, and friendly service, due largely to their extensive network of drivers and
their proprietary dispatching software named NorthRide.
Northwind Electric Cars drivers depend on frequent, automatic updates for the NorthRide mobile app. The
Northwind management team is concerned about unplanned system downtime and slow connection speeds
caused by high usage. Additionally, Northwind’s in-house data storage solution is unsustainable because of the
new influx of customer data that is retained. Data backups are made periodically on DVDs and stored onpremises at corporate headquarters.
Apps
NorthRide App
Northwind drivers use the NorthRide app to meet customer pickup requests. The app uses a GPS transponder
in each Northwind vehicle and Bing Maps APIs to monitor the location of each vehicle in the fleet in real time.
NorthRide allows Northwind dispatchers to optimize their driver coverage throughout the city.
When new customers call, the dispatcher enters their pickup locations into NorthRide. NorthRide identifies the
closest available driver. The dispatcher then contacts the driver with the pick-up details. This process usually
results in a pick-up time that is far faster than the industry average.
Drivers use NorthRide to track the number of miles they drive and the number of customers they transport.
Drivers also track their progress towards their established goals, which are measured by using key
performance indicators (KPIs).
NorthRide App 2.0
Northwind Electric Cars is growing quickly. New callers often wait for their calls to be answered because the
dispatchers are contacting their drivers to arrange pickups for other customers.
To support the growth of the business, Northwind’s development team completes an overhaul of the NorthRide
system that it has named NorthRide 2.0. When a dispatcher enters a customer’s pickup location, the address
and driving directions are automatically sent to the driver who is closest to the customer’s pickup location.
Drivers indicate their availability on the NorthRide mobile app and can view progress towards their KPI’s in real
time. Drivers can also record customer ratings and feedback for each pickup.
Business Requirements
Apps
NorthRideFinder App
Northwind Electric Cars needs a customer-facing website and mobile app that allows customers to schedule
pickups. Customers should also be able to create profiles that will help ensure the customer gets a ride faster
by storing customer information.
Predictor App
Northwind Electric Cars needs a new solution named Predictor. Predictor is an employee-facing mobile app.
The app predicts periods of high usage and popular pickup locations and provides various ways to view this
predictive data. Northwind uses this information to better distribute its drivers. Northwind wants to use the latest
Azure technology to create this solution.
Other Requirements
On-premises data must be constantly backed up.
Mobile data must be protected from loss, even if connectivity with the backend is lost.Dispatch offices need to have seamless access to both their primary data center and the applications and
services that are hosted in the Azure cloud.
Connectivity needs to be redundant to on-premises and cloud services, while providing a way for each
dispatch office to continue to operate even if one or all of the connection options fail.
The management team requires that operational data is accessible 24/7 from any office location.
Technical Requirements
Apps and Website
NorthRide / NorthRideFinder Apps:
The solution must support on-premises and Azure data storage.
The solution must scale as necessary based on the current number of concurrent users.
Customer pickup requests from NorthRideFinder must be asynchronous.
The customer pickup request system will be high in volume, and each request will have a short life span.
Data for NorthRideFinder must be protected during a loss of connectivity.
NorthRide users must authenticate to the company’s Azure Active Directory.
Northwind Public Website
The customer website must use a WebJob to process profile images into thumbnails
The customer website must be developed with lowest cost and difficulty in mind.
The customer website must automatically scale to minimize response times for customers.
Other Requirements
Data Storage:
The data storage must interface with an on-premises Microsoft SQL backend database.
A disaster recovery system needs to be in place for large amounts of data that will backup to Azure.
Backups must be fully automated and managed the Azure Management Portal.
The recovery system for company data must use a hybrid solution to back up both the on-premises
Microsoft SQL backend and any Azure storage.
Predictive Routing:
An Azure solution must be used for prediction systems.
Predictive analytics must be published as a web service and accessible by using the REST API.
Security:
The NorthRide app must use an additional level of authentication other than the employee’s password.
Access must be secured in NorthRide without opening a firewall port.
Company policy prohibits inbound connections from internet callers to the on-premises network.
Customer usernames in NorthRideFinder cannot exceed 10 characters.
Customer data in NorthRideFinder can be received only by the user ID that is associated with the data.

###EndCaseStudy###

You need to recommend the appropriate technology to provide the predictive analytics for passenger pickup.
What should you do?

A.
Use Power BI to analyze the traffic data and PowerPivot to categorize the results.

B.
Use HDInsight to analyze the traffic data and write a .NET program to categorize the results.

C.
Use Machine Learning Studio to create a predictive model and publish the results as a web service.

D.
Use Hadoop on-premises to analyze the traffic and produce a report that shows high traffic zones.

Explanation:
* Scenario: Predictive Routing:
/ An Azure solution must be used for prediction systems.
/ Predictive analytics must be published as a web service and accessible by using the REST API.
* Microsoft Azure Machine Learning Studio is a collaborative visual development environment that enables you
to build, test, and deploy predictive analytics solutions that operate on your data. The Machine Learning service
and development environment is cloud-based, provides compute resource and memory flexibility, and
eliminates setup and installation concerns because you work through your web browser.
What is Azure Machine Learning Studio?
https://azure.microsoft.com/en-us/documentation/articles/machine-learning-what-is-ml-studio/