Fundamentals of Big Data & Business Analytics

December 2023 Examination

1. There are various types of business analytics procedures that can be performed on a
problem statement. Assume you’re working with a low-cost airline based in India as an
analytics advisor. Define and state 3 different use cases of your choice that the airline
company could use. Mention briefly about how the different types of business analytics
procedures could be used in each of the use cases. (10 Marks)

Ans 1.

Introduction

In the contemporary era of digitalization, the aviation industry, like many others, stands at the
intersection of rapid technological developments and ever-increasing customer expectations.
Low-cost airlines, which focus primarily on reducing operating expenses to offer competitive
fares, operate under low profit margins. This makes efficient use of data and analytics critical
for success. With a flood of data, from customer preferences to flight operations, airlines have
a vast store of information at their disposal.

Concept & Applications

Business analytics is the process of using statistical methods, skills, technologies, and
practices to analyze historical data and gain new insights to improve strategic decision
making. This subset of data management solutions employs business intelligence and a variety
of methodologies such as data mining, statistical analysis, and predictive analytics. Various
business analytics types help analyze data and transform it into useful information, identifying
and predicting current trends and outcomes for smart, data-driven business decisions.
Certainly, a low-cost airline based in India can leverage business analytics in a variety
of ways to enhance its operations and decision-making. Here are three different use
cases for airlines:

1. Optimizing flight scheduling and route planning: Optimizing flight scheduling

and route planning is an important application of business analytics in the airline
industry, including low-cost airlines based in India. The process involves using data-
driven insights and mathematical models to create efficient flight schedules and
routes that maximize profitability, passenger satisfaction and operational
performance. Here's how this can be achieved through business analytics:

Data collection and analysis: Start by collecting historical data related to flight
schedules, passenger demand, and route performance. This includes data on flight
bookings, cancellations, on-time performance, load factors and historical route
profitability.

Demand forecasting: Use predictive analytics to forecast passenger demand for

different routes and flight times. Historical booking data, seasonality, and external
factors (e.g., holidays, special events) are considered to accurately predict future
demand.

Route Profitability Analysis: Perform overall profitability analysis for each route.
Calculate the costs associated with aircraft operation, fuel, maintenance, crew and
landing charges. Compare these costs to revenue projections to determine the
profitability of each route.

Use case: An airline aims to optimize its flight schedule and route planning to
increase operational efficiency and profitability..

Business analytics approach: Airlines can use predictive analytics to analyze

historical flight data, passenger demand, and external factors such as weather
conditions and holidays. Predictive analytics considers historical data trends to
determine the likelihood of particular future outcomes. It uses a number of
techniques such as data mining, machine learning algorithms, and statistical
modeling to predict the probability of events. It helps improve business areas
including customer service, efficiency, fraud detection and prevention, and risk
management. This allows airlines to increase the most profitable customers, improve
business operations, and determine customer responses and cross-sell
opportunities.

2. Enhancing Customer Experience and Loyalty: Enhancing customer experience

and loyalty through business analytics is important for any airline, including low-cost
airlines based in India. The ability to use data-driven insights to tailor services and
marketing efforts can significantly improve customer satisfaction and encourage
repeat business. Here's how this can be achieved:

Data collection and integration: Start by collecting and integrating customer data
from various touchpoints, including booking platforms, in-flight services, customer
feedback, and social media interactions. This data should include demographic
information, travel history, preferences and feedback.

Personalization: Apply predictive analytics to create personalized experiences for

different customer segments. Tailor offerings to meet the specific needs and
expectations of each customer group, such as in-flight services, seating preferences
and marketing promotions..

Recommendation systems: Implement recommendation systems that suggest

relevant travel options and promotions based on customers' preferences and past
behavior. Collaborative filtering and content-based recommendations can help
travelers find personalized travel experiences.

Use case: Airlines want to improve their customer experience and boost customer
loyalty to stay competitive.
Business analytics approach: Customer segmentation and sentiment analysis can
be applied to gain insight into traveler preferences and behavior. By identifying
different customer segments, the airline can tailor its services and marketing
strategies to meet the specific needs and expectations of each group.
Recommendation systems and personalized marketing campaigns can also be
developed to provide customers with personalized travel options, promotions and
rewards, increasing customer satisfaction and loyalty.

3. Maintenance Predictive Analytics for Fleet Management: Maintenance

predictive analytics for fleet management is an important application of business
analytics in the airline industry, ensuring that aircraft remain in optimal operating
condition and minimizing disruptions. Here's how maintenance predictive analytics
can be used to enhance fleet management for a low-cost airline based in India:

Data collection and sensor integration: Apply IoT sensors and data collection
systems on the aircraft to monitor various components in real time. These sensors
capture data on engine performance, fuel consumption, temperature, pressure and
other relevant parameters. Data is transmitted to a central database .

Data preprocessing and quality control: Preprocess the collected data to clean it and ensure
its quality. This includes removing outliers, handling missing values, and normalizing the data
for analysis.

Use case: Airlines strive to minimize aircraft downtime, reduce maintenance costs, and
enhance overall fleet management.

Business Analytics Approach: Implementing predictive maintenance processes, including IoT

sensors on aircraft, allows the airline to collect real-time data on the performance of aircraft
components. Predictive maintenance models can then analyze this data to predict when
maintenance will be needed. Machine learning algorithms can be used to detect anomalies and
identify potential issues before they lead to equipment failures. This proactive approach
reduces unplanned maintenance and improves flight reliability. Additionally, prescriptive
analysis can suggest optimal maintenance schedules considering factors such as crew
availability, aircraft utilization and routes, ensuring that maintenance occurs at the least
disruptive time while maximizing aircraft availability.

Conclusion

These use cases demonstrate the versatility of business analytics in helping a low-cost airline
based in India make informed decisions, improve operational efficiency, increase customer
satisfaction and reduce costs. By leveraging data-driven insights, airlines can gain a
competitive edge in the highly competitive aviation industry.

2. Alset is a multi-national top end electric vehicle manufacturer and a market leader in
providing autonomous driving features. No other car manufacturer in the world is
anywhere near Alset in the autonomous driving space. Being at the forefront comes with
its challenges.

Alset’s top management heard about its software not being able to differentiate
between 2-wheelers, i.e., motorcycles and bicycles. Which caused an erroneous speed
estimation of those 2-wheelers and led to some non-fatal accidents. The organization
wants to avoid this. For which, it needs a lot of data to train the model. But it’s not
storing any data from their cars. Some of the users agreed to share the image/video data
from their cars but Alset doesn’t have a data collection or storage or processing
methodology in place because all the processing thus far has been on the on-board units
(OBUs). To solve this the firm consults and gives you the responsibility to design their
pipeline. They don’t have any historical data on any kind of 2-wheelers.

How do you approach this problem holistically using both big data and business analytics
concepts? Explain the big data component for storage and processing, and analytics
component of approaching this use case and what would you do in each of the stages.
Goal is to help company collect, store, process data and better predict and classify
between different kinds of 2-wheelers. (10 Marks)
Ans 2.

Introduction

In the rapidly evolving automotive industry, where autonomous vehicles have gained
significant popularity, the role of data analytics has become increasingly indispensable.
Alsace, the leading power in electric vehicles and autonomous driving, finds itself at a turning
point. The recent revelation that their autonomous software struggles to distinguish between
motorcycles and bicycles underlines the need for a robust big data infrastructure. The
implications are more than just technical; They directly impact user safety, brand reputation
and market leadership.

Concept & Application

To holistically understand the problem facing Alsat, which includes improving interoperability
between 2-wheel vehicles (motorcycles and bicycles) in their autonomous driving system, we
can leverage big data and business analytics concepts . It aims to establish data collection,
storage and processing methodology and develop an analytical framework that helps in
accurately predicting and classifying different types of 2-wheelers. Here's a step-by-step
approach to address this challenge:

Phase 1: Data Collection

Big Data Component for Data Collection:

• Implement data streaming: Install data streaming capabilities on Alsat's vehicles to collect
real-time data, including images and video, from sensors and cameras. This data must be sent
to a centralized platform for processing.

• User data sharing: Encourage users who have agreed to share data from their vehicles to
upload image and video data to the centralized platform.
Analytics Component in Data Collection:

• Data collection strategy: Define a clear strategy for data collection, including the type of
data to be collected (images, videos), data sources (vehicles and users), and frequency of data
collection.

• Data quality control: Implement data quality checks to ensure that the data collected is
consistent, reliable, and free from errors.

Phase 2: Data Storage

Big Data Component for Data Storage:

• Scalable Storage Infrastructure: Establish a scalable and cost-effective data storage

infrastructure that can handle the flow of large amounts of data. Consider cloud-based storage
solutions for flexibility and scalability.

Analytics Component in Data Storage:

• Data management: Develop a data management strategy that includes data classification,
tagging, and indexing to facilitate data retrieval and analysis.

• Data security and privacy: Implement strong security measures to protect sensitive user data.
Ensure compliance with data privacy regulations, obtain necessary user consent for data use.

Phase 3: Data Processing

Big Data Component for Data Processing:

• Distributed processing: Employ distributed computing frameworks and technologies (e.g.,

Hadoop, Spark) to efficiently process large datasets.
• Real-time and batch processing: Apply both real-time and batch data processing to handle
incoming data streams and historical data.

Analytics Component in Data Processing:

• Data pre-processing: Clean, pre-process, and transform data as needed for machine learning
and analytics, including resizing images, video frame extraction, and feature extraction.

• Machine Learning Models: Develop and train machine learning models to accurately classify
two-wheelers. Include deep learning models for image and video analysis.

Phase 4: Data Analysis and Predictive Modeling

Big Data Component for Data Analysis:

• Data Visualization: Use data visualization tools and techniques to gain insights from the
processed data. Visualization can help identify patterns and anomalies in the behavior of two-
wheelers.

Analytics Component in Data Analysis:

• Classification and Prediction: Develop predictive models that can accurately classify and
predict the type of 2-wheeler vehicles (motorcycle or bicycle). Continuous model refinement
based on ongoing data analysis is important.

Phase 5: Implementation and Feedback Loop

Big Data Component for Implementation:

• Integration with autonomous driving systems: Integrate predictive models and
classification algorithms into Alset's autonomous driving systems to improve differentiation
between two-wheelers.

Analytics Component in Implementation:

• Feedback mechanisms: Establish feedback loops to continually refine models and

algorithms based on real-world performance. Collect data on system accuracy and user
feedback.

Phase 6: Deployment

Big Data Component for Deployment:

• Scalable infrastructure: Ensure that data storage and processing infrastructure can scale with
growing datasets and evolving analytics needs..

Analytics Component in Deployment:

• Monitoring and maintenance: Implement continuous monitoring of system performance, and

provide maintenance and updates to the classification model as needed.

Phase 7: Continuous Improvement

Feedback Loop:

- Establish feedback loops for continuous improvement of the classification model. Collect
data on system accuracy and user feedback. Incorporate this feedback into model refinement.

Model Updates:
Regularly update classification models to adapt to changing road conditions, new types of two-
wheelers and evolving user data.

Phase 8: Deployment

Scalable Infrastructure:

Ensure that data storage and processing infrastructure can scale to accommodate growing
datasets and the demands of autonomous driving systems.

Monitoring and Maintenance:

Implement continuous monitoring of system performance and provide regular maintenance and
updates to the classification model.

Conclusion

By following this holistic approach, Alset can solve the challenge of accurately distinguishing
between two-wheelers, preventing wrong speed estimation and improving overall safety. By
following this step-by-step approach, Alset can build an effective pipeline for collecting,
processing, and analyzing data to improve the interoperability between motorcycles and
bicycles in its autonomous driving system. The use of big data concepts enables efficient data
collection and processing, while business analytics concepts help in model development and
continuous improvement based on real-world feedback and analysis.

3.Costco, a high-volume low-cost retailer like D-Mart is planning to start operations in

India. Help them and identify 10 locations, which will have high reachability, multi city
presence and many more that aid Costco in building a pan India chain in 5-10 years. You
are not just helping them identify the top cities but also the localities and the specific
area in which they should open.
a. Which type of analysis could be used to solve this problem and maximise the outcome
for Costco? Explain the data needed to solve this problem. (5 Marks)

Ans 3a.

Introduction

Business analytics plays a vital role in transforming raw data into valuable insights that can
inform decision making. By using business analytics tools, organizations can gain a deeper
understanding of the primary and secondary data coming out of their activities, enabling them
to refine their processes and improve productivity. To maintain a competitive edge, businesses
need to stay ahead of their peers and leverage the latest tools to improve efficiency and
generate more profits.

Concept & Application

A combination of business analytics and data-driven approaches are being used to help Costco
identify 10 locations for high-reach, multi-city presence in India and build a pan-India chain
within 5-10 years.Here's a step-by-step guide on how to deal with this problem:

Type of Analysis:

Market Potential and Location Analysis

1. Market Research: Conduct extensive market research to understand the retail landscape,
consumer behavior and market trends in India. This will include data on consumer
demographics, preferences and purchasing power.

2. Data collection: Gather data on potential locations, including information on population

density, income levels, and competition (e.g., existing retail stores, local brands).
3. Geospatial Analytics: Use geospatial analytics to map and analyze data geographically.
Geographic information systems (GIS) can help understand population distribution,
transportation networks, and urban development.

4. Demographic analysis: Analyze demographic data to identify cities and specific

neighborhoods with high concentrations of the target customer base (for example, middle- to
upper-middle income households).

5. Economic data: Collect economic data related to the cost of living, disposable income and
purchasing power of consumers in different regions.

6. Infrastructure and accessibility: Evaluate the accessibility and infrastructure of potential

locations. This includes proximity to highways, public transportation, and other logistical
considerations.

7. Competition Analysis: Assess the competitive landscape in each region to understand the
presence of similar retailers and the possibility of market saturation.

8. Local Preferences: Collect data on local consumer preferences, including purchasing habits
and cultural nuances that can impact a retail store's success.

9. Future growth prospects: Consider data on urbanization trends and expected population
growth to select locations with long-term potential.

10. Regulatory and legal considerations: Ensure compliance with local regulations, zoning
laws, and licensing requirements.

Data Needed:

To perform the above analysis, Costco would need access to a wide range of data, which
may include:
• Demographic data: population size, age distribution, income levels and household size.

• Economic data: information on gross domestic product, disposable income and cost of living.

• Geospatial data: geographic coordinates, urban development data and transportation

networks.

• Retail data: Data on existing retail stores, local brands and their performance.

• Consumer behavior data: Surveys and studies on consumer preferences and purchasing
habits.

• Competition data: Data on competitors, their market share and their locations.

• Infrastructure data: information on transport networks, road quality and accessibility.

• Urbanization and growth projections: data on urbanization rates and future population growth
projections.

Conclusion

By analyzing this data, Costco can make informed decisions on the most suitable locations to
set up its retail chain in India. It can ensure high reach, multi-city presence and long-term
success by selecting locations that align with its target customer base and the dynamics of the
emerging Indian market.

b. For the purposes of customer relationship management, Costco is planning to collect

data of customers and their purchases to provide better services. Which kind of data
storage should they prefer? An on-premises data facility or a cloud data facility? Why?
(5 Marks)

Ans 3b.
Introduction

CRM helps businesses build relationships with their customers which in turn builds loyalty
and customer retention. Since customer loyalty and revenue are both attributes that affect a
company's revenue, CRM is a management strategy that results in increased profits for a
business. At its core, a CRM tool creates a simple user interface for the collection of data that
helps businesses identify and communicate with customers in a scalable manner.

Concept & Application

For customer relationship management (CRM) purposes, the choice between on-premises data
storage and a cloud data facility depends on a variety of factors, including Costco's specific
needs and resources. Each option has its own advantages and considerations:

On-Premises Data Facility:

Pros:

1. Control: Costco will have complete control over its data, including data security, access,
and customization of hardware and software configuration.

2. Compliance: It may be easier to comply with data privacy regulations and industry-specific
requirements by maintaining data on-premises, as it provides more direct control over data
management.

3. Data Sensitivity: If Costco deals with highly sensitive customer data, such as personally
identifiable information (PII), keeping it on premises can provide an additional layer of
security and reduce the risk of data breaches.

Cons:
1. Infrastructure costs: Setting up and maintaining on-premises data facilities can be
expensive, including hardware, software, IT staff, and maintenance expenses.

2. Scalability: Expanding storage and processing capacity may require significant upfront
investment and may take longer to deploy.

3. Disaster Recovery: Implementing effective disaster recovery and backup systems can be
complex and expensive.

Cloud Data Facility:

Pros:

1. Cost-effective: Cloud data facilities generally offer a cost-effective pay-as-you-go model,

which minimizes upfront capital expenditure.

2. Scalability: Cloud providers provide scalable solutions, allowing Costco to quickly adapt to
changing data storage needs.

3. Access: Data can be accessed from anywhere, facilitating remote access for employees and
enabling seamless collaboration.

4. Managed Services: Cloud providers often offer managed services, reducing the burden of
maintenance, updates, and security on Costco's IT team.

5. High availability: Cloud providers typically have redundant systems and robust disaster
recovery mechanisms.

Cons:
1. Data Security: While cloud providers make significant investments in security, some
organizations may have concerns about data security in the cloud. However, these concerns
can often be mitigated with proper configuration and encryption.

2. Data Privacy: Compliance with data privacy regulations may require careful management of
data stored in the cloud, which may include data encryption and data residency considerations.

Recommendation: Given the large-scale operations and customer data involved in a retailer
like Costco, a cloud data feature is often a practical and cost-effective option for CRM data
storage. It provides scalability, accessibility, and managed services that can effectively
support Costco's CRM goals. However, data security and compliance are paramount, and
Costco should work closely with its cloud provider to ensure appropriate security measures are
in place and data privacy regulations are followed.

Conclusion

The specific choice between on-premises and cloud data storage should be made based on
Costco's specific business needs, budget, and risk tolerance. It is common for organizations to
use a combination of both options, known as a hybrid cloud approach, to balance control and
scalability.