Chapter-9: Big Data, Cloud Computing, and Location Analytics: Concepts and Tools

1. DEFINITION OF BIG DATA and The “V”s That Define Big Data
Big Data
Big Data refers to the enormous volumes of data that traditional data processing methods find challenging to handle.
This term has gained substantial attention in recent years, drawing interest from business users and IT professionals
alike. Big Data's appeal lies in its potential to provide deep insights into customers, clients, and business operations,
fostering growth and profitability. However, as more data becomes available in various forms and at rapid speeds,
processing this data timely and effectively with conventional tools becomes increasingly impractical. This
phenomenon, known as Big Data, is often overhyped, leading some experts to consider it an overused buzzword.
Essentially, Big Data is a relative term—its definition varies based on an organization's size and the specific context in
which it is used. Traditionally, Big Data described the massive datasets handled by large organizations like Google or
research projects at NASA. For most businesses, however, "big" depends on their scale and the extent to which they
push data analytics to uncover new insights and opportunities.
Big Data encompasses more than just large volumes of data; it includes various characteristics that make it complex to
manage and analyze. The primary sources of Big Data are diverse and extensive, ranging from web logs, RFID, GPS,
and sensor networks to social media, scientific research, and e-commerce practices. While storing and analyzing large
volumes of data isn't new—dating back to the advent of data warehouses in the early 1990s—the scale and complexity
of modern Big Data have escalated. Organizations now seek to store and analyze more detailed transaction data, web-
and machine-generated data, to gain a deeper understanding of customer behavior and business drivers. This
continuous growth in data volume presents both challenges and opportunities for businesses, prompting the
development of new technologies to manage and process Big Data effectively.
The "V"s That Define Big Data
Big Data is typically defined by three core characteristics: volume, variety, and velocity. However, leading solution
providers have introduced additional "V"s, including veracity, variability, and value proposition, to encapsulate the full
spectrum of Big Data's attributes.
Volume is the most prominent trait of Big Data, referring to the massive amounts of data generated from various
sources. The exponential increase in data volume is driven by factors like transaction-based data, social media, and
sensor data. While excessive data volume once posed significant storage challenges, advances in technology and
decreasing storage costs have shifted the focus to determining the relevance of data and deriving value from it. The
scale of data has grown from terabytes (TB) to zettabytes (ZB), reflecting the staggering increase in data generated
over the years.
Variety pertains to the different types of data formats, from traditional structured data in databases to unstructured
data like text documents, emails, videos, and social media posts. Estimates suggest that 80 to 85% of organizational
data is unstructured or semi-structured, which traditional database schemas cannot easily manage. Despite this,
unstructured data holds significant value and must be included in analyses to support comprehensive decision-making.
Velocity involves the speed at which data is produced and processed. With the rise of technologies like RFID tags,
sensors, and GPS devices, there is an increasing need to handle torrents of data in near real-time. Velocity is often
overlooked but is critical for organizations that need to react quickly to data as it is generated. In time-sensitive
environments, the value of data diminishes rapidly, making timely access and analysis crucial for beneficial outcomes.
Data stream analytics, or "in-motion analytics," has emerged to address the need for rapid data processing and
reaction.
Veracity, a term popularized by IBM, refers to the accuracy, quality, and trustworthiness of data. High-quality, reliable
data is essential for generating meaningful insights. Tools and techniques are used to ensure data veracity,
transforming raw data into valuable and trustworthy information.
Variability addresses the inconsistencies and variations in data flow. Data can be highly inconsistent, with peaks
triggered by events, trends, or seasonal changes. Managing these variations, especially with the involvement of social
media, presents significant challenges.
Value Proposition is the most exciting aspect of Big Data, emphasizing its potential to uncover patterns and
anomalies that provide significant business insights. Analyzing large, feature-rich datasets can reveal more profound
insights and drive better decisions, which is a critical need for organizations. Big Data analytics aims to harness this
value, offering greater business insights and improved decision-making capabilities.
In conclusion, Big Data is a multifaceted concept characterized by its volume, variety, velocity, veracity, variability,
and value proposition. Understanding and leveraging these characteristics allows organizations to derive significant
insights and competitive advantages from their data, making Big Data an essential component of modern business
strategy.

2. Critical Success Factors for Big Data Analytics

To successfully implement Big Data analytics, organizations must consider several critical success factors. These
factors ensure that the investment in Big Data delivers significant business value and supports organizational
objectives. Here are the detailed explanations of the critical success factors:
1. Clear Business Need
o Alignment with Vision and Strategy: Big Data analytics should be driven by a clear business need,
ensuring that the investment supports the overall vision and strategic goals of the organization. This
means identifying specific business problems or opportunities that Big Data analytics can address,
such as improving customer insights, enhancing operational efficiency, or driving innovation.
o Strategic, Tactical, and Operational Levels: The need for Big Data analytics should be evident at all
levels of the organization. Whether it’s for strategic decision-making, tactical planning, or day-to-day
operations, the data-driven insights should provide measurable benefits and drive business growth.
2. Strong, Committed Sponsorship
o Executive Champion: The success of Big Data initiatives heavily relies on having strong and
committed sponsorship from top executives. This executive champion should actively support and
advocate for the Big Data projects, ensuring that they receive the necessary resources, attention, and
organizational buy-in.
o Organization-Wide Support: For enterprise-wide transformations, sponsorship needs to extend
beyond individual departments to encompass the entire organization. This helps in overcoming
resistance, aligning different units, and ensuring cohesive efforts towards leveraging Big Data.
3. Alignment Between Business and IT Strategy
o Support Business Strategy: The Big Data analytics efforts should always support and enhance the
business strategy. This alignment ensures that analytics initiatives are focused on delivering business
value rather than being driven by technology for its own sake.
o Enabling Role of Analytics: Analytics should be seen as an enabler for executing the business
strategy effectively. This involves close collaboration between business and IT teams to ensure that
the data infrastructure, tools, and techniques are aligned with business needs.
4. Fact-Based Decision-Making Culture
o Data-Driven Decisions: Creating a culture where decisions are based on data rather than intuition or
gut feeling is crucial. This involves fostering a mindset that values empirical evidence and analytical
rigor in the decision-making process.
o Experimentation and Adaptation: A fact-based culture encourages experimentation to determine
what works and what doesn’t. This iterative approach helps in continuously refining strategies and
improving outcomes based on data-driven insights.
 o Senior Management Role: Senior management plays a key role in promoting a fact-based decision-
making culture by:
▪ Recognizing that some individuals may resist change.
▪ Being vocal supporters of data-driven initiatives.
▪ Emphasizing the discontinuation of outdated methods.
▪ Insisting on seeing the analytics behind decisions.
▪ Linking incentives and compensation to desired behaviors and outcomes.
5. Strong Data Infrastructure
o Data Warehouses and Beyond: While traditional data warehouses have been the backbone of data
infrastructure, the era of Big Data requires integrating these with new technologies to handle the
increased volume, variety, and velocity of data.
o Holistic Infrastructure: Success in Big Data analytics necessitates a holistic infrastructure that
combines old and new technologies seamlessly. This ensures efficient data capture, storage,
processing, and analysis, enabling organizations to derive meaningful insights from their data.
Additional Factors
6. High-Performance Computing Techniques
o In-Memory Analytics: Processing data in-memory across multiple nodes allows for near real-time
insights, making it possible to solve complex problems quickly and accurately.
o In-Database Analytics: Performing analytics within the database reduces data movement and
enhances data governance, speeding up the time to insights.
o Grid Computing: Utilizing a shared pool of IT resources promotes efficiency and lowers costs,
offering better performance for Big Data processing.
o Appliances: Combining hardware and software in scalable units provides fast and efficient data
processing capabilities.
7. Addressing Challenges
o Data Volume: Organizations must have the capability to capture, store, and process vast amounts of
data quickly to ensure that decision-makers have access to the latest information.
o Data Integration: Efficiently combining data from various sources and structures is crucial for
comprehensive analysis.
o Processing Capabilities: Real-time data processing (stream analytics) is essential to leverage the
most value from rapidly arriving data.
o Data Governance: Maintaining robust data governance practices to address security, privacy,
ownership, and quality issues is critical as the volume and variety of data increase.
o Skills Availability: There is a growing need for skilled data scientists who can harness new tools and
methodologies to analyze Big Data effectively.
o Solution Cost: Ensuring a positive return on investment by managing the costs associated with Big
Data solutions is essential. This involves balancing experimentation with practical applications that
deliver tangible business value.
By focusing on these critical success factors, organizations can effectively navigate the complexities of Big Data
analytics and unlock its full potential to drive business success.

3. Big Data Technologies Overview

Big Data technologies enable the processing and analysis of massive, complex datasets that traditional data-processing
software cannot handle efficiently. These technologies often leverage commodity hardware, parallel processing,
nonrelational data storage, and advanced analytics. The three primary technologies discussed here are MapReduce,
Hadoop, and NoSQL databases.
MapReduce
MapReduce, pioneered by Google and popularized through its seminal paper by Dean and Ghemawat in 2004, is a
distributed programming model and implementation for processing large datasets across clusters of commodity
hardware. The fundamental idea behind MapReduce is to break down complex data processing tasks into smaller,
manageable units that can be executed in parallel across numerous nodes in a cluster.
Key Characteristics:
• Parallel Processing: Enables significant scalability by distributing tasks across multiple nodes, each handling
a portion of the dataset concurrently.
• Functional Programming Model: Programs written in MapReduce are expressed in a functional style,
consisting of two main phases:
 o Map: Processes input data and generates intermediate key-value pairs.
o Reduce: Aggregates and processes the intermediate data to produce the final output.
• Automatic Parallelization: Simplifies the task of parallel and distributed computing for programmers,
abstracting away the complexities of managing distributed systems.
Applications: MapReduce finds applications in various domains, including:
• Data Indexing and Search: Efficiently indexing vast amounts of data for quick retrieval.
• Text Analysis: Analyzing and processing large corpora of text data for insights and patterns.
• Machine Learning: Training models on large datasets distributed across clusters.
• Graph Analysis: Analyzing relationships and networks within large datasets.
Advantages: MapReduce facilitates the processing and analysis of massive datasets that traditional databases and
single-machine processing cannot handle efficiently. Its scalability, fault tolerance, and ability to handle diverse data
types make it indispensable for Big Data analytics.
Hadoop
Hadoop is an open-source framework designed for distributed storage and processing of large datasets across clusters
of commodity hardware. Originating from Doug Cutting's efforts at Yahoo! based on Google's MapReduce concept,
Hadoop has evolved into a robust ecosystem supported by the Apache Software Foundation.
Core Components:
• Hadoop Distributed File System (HDFS): A scalable and fault-tolerant distributed storage system that spans
all nodes in a Hadoop cluster. It stores data in a distributed manner, ensuring redundancy and availability.
• MapReduce Framework: Executes data processing tasks across the nodes in the Hadoop cluster, utilizing the
data stored in HDFS.
• Job Tracker and Name Node: Manage and coordinate jobs and data location information within the Hadoop
cluster, ensuring efficient task execution and fault tolerance.
Workflow:
1. Data Ingestion: Unstructured and semistructured data from various sources (e.g., log files, social media
feeds) is ingested into Hadoop clusters.
2. Processing: Users submit MapReduce jobs, which are executed across the cluster. Each node processes its
portion of the data in parallel.
3. Data Storage: Processed data is stored back into HDFS, where it remains available for further analysis or
export to other systems.
4. Analysis: Data scientists and analysts can further analyze the processed data using various tools and
frameworks integrated with Hadoop.
Advantages: Hadoop's scalability, fault tolerance, and cost-effectiveness (due to its use of commodity hardware)
make it ideal for organizations dealing with vast amounts of data. It supports batch processing and serves as a
foundational platform for building data-driven applications and analytics solutions.
NoSQL Databases
NoSQL (Not Only SQL) databases represent a category of database management systems designed for handling large
volumes of unstructured and semistructured data. Unlike traditional relational databases, NoSQL databases offer
flexible schema design and horizontal scalability.
Key Features:
• Schema Flexibility: NoSQL databases accommodate diverse data types and structures without requiring a
predefined schema, making them suitable for agile development and managing evolving data models.
 • Horizontal Scalability: Scales out across multiple nodes, allowing seamless expansion of storage and
processing capabilities by adding more servers to the cluster.
• High Performance: Optimized for high read and write throughput, crucial for applications requiring rapid
access to large datasets.
• Types: NoSQL databases include various types such as key-value stores, document stores, column-family
stores, and graph databases, each tailored to specific data management and access patterns.
Applications:
• Real-time Web Applications: Handling high volumes of user-generated data and serving dynamic content.
• IoT Data Management: Storing and analyzing data streams from connected devices in real time.
• Content Management Systems: Managing and delivering diverse content types efficiently.
• Analytics and Personalization: Powering applications that require quick access to customer behavior and
interaction data.
Challenges and Evolution: NoSQL databases often sacrifice ACID compliance for scalability and performance.
However, ongoing developments within the open-source community and efforts by commercial vendors aim to
address these limitations, improving management tools and enhancing transactional capabilities.
Conclusion
These technologies—MapReduce, Hadoop, and NoSQL databases—are pivotal in the realm of Big Data analytics,
each addressing specific challenges in processing, storing, and analyzing massive datasets. Together, they empower
organizations to derive actionable insights, optimize operations, and innovate across various industries, leveraging the
power of distributed computing and flexible data management approaches. Understanding their capabilities and
applications is essential for businesses seeking to harness the full potential of Big Data for strategic advantage.

4. Use case of Hadoop and Data Warehouse

The advent of Big Data has significantly influenced data warehousing, previously the cornerstone of decision support
technologies. While enterprise data warehouses, primarily using RDBMS (Relational Database Management
Systems), have been central to decision-making, they now share the stage with Big Data technologies like Hadoop.
This shift raises questions about whether Big Data will replace traditional data warehousing. However, this is not a
realistic perspective as both technologies serve different purposes and complement each other rather than compete.
Use Cases for Hadoop
Hadoop is particularly suitable for managing large volumes of unstructured or semi-structured data that arrives at high
velocity. Here are the primary use cases for Hadoop:
1. Hadoop as the Repository and Refinery
Hadoop serves as a cost-effective and flexible repository for capturing and refining raw data.
• Data Ingestion: Hadoop can efficiently capture vast amounts of data from diverse sources such as sensors,
social media, and clickstreams. This is made possible by the Hadoop Distributed File System (HDFS), which
allows data to be spread across multiple nodes, facilitating parallel processing and storage.
• Data Storage: Unlike traditional storage systems that may rely on expensive hardware, Hadoop leverages
commodity hardware, reducing the cost per gigabyte of storage. Data is stored locally on each node, which
often results in faster data access compared to storage area networks.
• Data Refinement: Once captured, raw data needs to be processed and transformed into a more structured
format. Hadoop excels in data transformation tasks, using tools like MapReduce or newer processing engines
like Apache Spark. These tools can handle complex data formats, such as clickstream data or sensor logs,
efficiently converting them into structured data sets.
2. Hadoop as the Active Archive
Hadoop provides a robust and accessible long-term storage solution for historical data.
• Economic Storage: Hard disks in Hadoop clusters can be treated similarly to magnetic tapes for long-term
storage but with better accessibility. This approach is economically feasible due to the declining costs of disk
storage and the use of commodity hardware.
• Data Retrieval: Unlike tape storage, which can be slow and prone to degradation, data stored in Hadoop can
be quickly accessed and analyzed. This makes Hadoop a valuable tool for organizations that need to revisit
historical data for new analyses or to meet changing business contexts.
• Long-Term Data Value: Keeping data accessible on Hadoop enables organizations to derive ongoing value
from it. For example, raw data from clickstreams can be re-analyzed as new business questions arise,
providing deeper insights over time.
Use Cases for Data Warehousing
Data warehouses are designed to handle structured data and support complex queries and business intelligence (BI)
tools. Here are the primary use cases for data warehouses:
1. Data Warehouse Performance
Data warehouses offer high performance for complex queries through advanced optimization techniques.
• Indexing and Materialized Views: Data warehouses use various indexing methods and materialized views to
speed up query response times. Indexes help locate data quickly, while materialized views store pre-computed
results for complex queries.
• Cost-Based Optimization: A key feature of data warehouses is the cost-based optimizer, which analyzes SQL
queries and determines the most efficient execution plan based on the database's schema and data statistics.
This ensures queries are executed as quickly as possible, optimizing resource use.
2. Integrating Data that Provides Business Value
The integration of data from multiple sources into a single repository is a core function of data warehouses.
• Data Integration: Data warehouses pull data from various subject areas and applications, creating a unified
repository. This involves data modeling and ETL (Extract, Transform, Load) processes to clean and
standardize data, ensuring it is consistent and reliable.
• Business Insights: By integrating data, data warehouses enable organizations to answer critical business
questions and derive actionable insights. For example, a data warehouse might integrate sales, customer, and
inventory data to provide a comprehensive view of business performance.
3. Interactive BI Tools
Data warehouses support interactive BI tools that empower business users to analyze data and generate reports.
• Self-Service BI: Tools like MicroStrategy, Tableau, and IBM Cognos allow business users to create and
modify reports without relying on IT staff. This democratizes data access, enabling more employees to
leverage data in their decision-making processes.
• Enhanced Decision-Making: The interactivity provided by BI tools fosters a deeper understanding of the
business. Users can quickly adjust their analyses to explore different aspects of the data, leading to more
informed and strategic decisions.
Complementary Roles
While Hadoop and data warehouses have distinct strengths, they can be used together to create a more robust data
ecosystem. Hadoop can handle the ingestion and initial processing of large, unstructured data sets, while data
warehouses can integrate and analyze structured data to provide business insights. This complementary approach
leverages the strengths of both technologies, ensuring that organizations can manage and utilize data effectively across
its lifecycle.
5. In the era of Big Data, are we about to witness the end of data warehousing?
No, we are not witnessing the end of data warehousing. Instead, data warehousing and Big Data technologies like
Hadoop are complementary and serve different purposes.
Data Warehousing Strengths
Data warehousing excels at managing and integrating structured data, providing robust business intelligence (BI)
capabilities through seamless integration with BI tools for sophisticated reporting and analysis. It also offers advanced
indexing and query optimization, ensuring real-time analytics and efficient performance.
Big Data and Hadoop Advantages
Big Data technologies like Hadoop are designed for handling large volumes of unstructured and semi-structured data.
They excel in scalability, leveraging commodity hardware to scale horizontally, and offer cost-effective storage
solutions for extensive datasets. Hadoop is particularly useful for capturing and processing raw data, as well as
providing economical long-term storage for historical data.
Complementary Technologies
In practice, these technologies are used in complementary ways. Hadoop is often employed for the pre-processing and
archival of raw data, while data warehousing is utilized for integrated analytics and structured data analysis. Hybrid
architectures that combine both technologies are becoming increasingly common, leveraging Hadoop for raw data
processing and data warehousing for integrated analytics.
Conclusion
Rather than replacing data warehousing, Big Data technologies enhance overall data management and analytics
capabilities, addressing different aspects of the data lifecycle and working together to provide comprehensive
solutions.

6. Coexistence of Hadoop and Data Warehouse

There are several scenarios where a combination of Hadoop and relational DBMS-based data warehousing
technologies is beneficial:
1. Storage and Archiving: Use Hadoop for storing and archiving multistructured data. Connectors to relational
DBMS can extract required data for analysis, using SQL-MapReduce functions for efficient data loading
between HDFS and the database.
2. Data Processing: Use Hadoop for filtering, transforming, and consolidating multistructured data. Connectors
can extract processed results to the relational DBMS for further analysis.
3. Analytics Platform: Use Hadoop to analyze large volumes of multistructured data and publish results, which
can then be integrated back into traditional data warehousing environments or shared data stores.
4. Investigative Computing: Use a relational DBMS with MapReduce capabilities to analyze a combination of
structured and multistructured data, leveraging both SQL and MapReduce functions.
5. Front-End Query Tool: Employ a query tool to access and analyze data stored in both Hadoop and relational
DBMS.
These scenarios demonstrate the use of connectivity software to exchange data between Hadoop and relational
DBMSs. The industry is moving toward tighter integration of these technologies, which will streamline systems,
reduce data movement, provide a unified metadata store, and offer a single interface for users and analytical tools. The
ultimate goal is to derive valuable insights from data, regardless of its storage format.

7. Big Data and Stream Analytics

One key characteristic of Big Data is velocity, referring to the rapid creation and streaming of data into analytics
environments. Traditional analytics methods, which work with static data, often fail to make timely or contextually
accurate decisions. Therefore, stream analytics has become essential for processing and analyzing data as soon as it is
created or streamed.
Challenges of Data Storage
While recording all data for potential future use is important, the increasing number of data sources makes this
approach less feasible. Current storage technology cannot keep up with the volume of digital information generated.
Real-time detection of changes and pattern variations is critical, leading to the need for stream analytics.
Stream Analytics
Stream analytics, or data-in-motion analytics, involves extracting actionable information from continuous data flows,
known as streams, which are sequences of data elements or tuples. If a single tuple is not sufficient for analysis, a
window of data, consisting of a sequence of tuples, is used. Stream analytics is gaining popularity due to decreasing
time-to-action and technological advancements that enable real-time data capture and processing.
Applications of Stream Analytics
• Energy Industry: Smart grids use real-time data to optimize power distribution and predict short-term energy
demands.
• e-Commerce: Companies like Amazon and eBay analyze real-time data from customer interactions to
enhance shopping experiences and increase sales.
• Telecommunications: Analyzing call detail records and social media data helps prevent churn and optimize
marketing campaigns.
 • Law Enforcement and Cybersecurity: Streaming data enhances crime prevention, cybersecurity, and real-
time situational awareness.
• Power Industry: Smart meters and sensors provide real-time data for optimizing power supply and
integrating weather data for better energy forecasts.
• Financial Services: Real-time analysis of trading data enables faster buy/sell decisions and fraud detection.
• Health Sciences: Streaming diagnostic data from medical devices improves patient care and safety.
• Government: Real-time analytics help manage natural disasters, monitor environmental conditions, and
optimize traffic flow in congested cities.
Stream Analytics vs. Perpetual Analytics
While both terms are often used interchangeably, there is a distinction. Stream analytics applies transaction-level logic
to real-time observations within a specified window size. Perpetual analytics evaluates each new observation against
all prior observations without a window size, offering deeper real-time insights. Each approach has its advantages
depending on the application's requirements.
8. CLOUD COMPUTING AND BUSINESS ANALYTICS
Cloud computing is a transformative technology that provides on-demand network access to shared computing
resources like networks, servers, storage, and applications. It allows users to leverage scalable and virtualized
resources over the Internet without needing to understand or manage the underlying infrastructure. Cloud computing
has evolved from earlier trends like utility computing and SaaS, and it includes several service models:

Models of Cloud Computing

1. Data as a Service (DaaS): Data as a Service (DaaS) allows businesses to access and manage data irrespective
of its physical location. Whether the data resides on local servers or in a remote cloud environment, DaaS
provides seamless access through standard protocols like SQL and XML. This model centralizes data quality
management, ensuring consistency and accuracy across different systems and applications. With DaaS,
businesses benefit from improved data agility, as changes to data structures or location-specific requirements
can be implemented easily. Additionally, DaaS enhances data quality by consolidating updates to a single
point, making it a highly efficient model for data management and integration.
2. Software as a Service (SaaS): Software as a Service (SaaS) delivers software applications over the Internet,
eliminating the need for users to install and maintain the software on their local devices. Users can access
these applications through web browsers or mobile apps, with the cloud provider managing the underlying
infrastructure and software updates. This model significantly reduces the burden of software maintenance and
IT management for consumers, as they only need to focus on using the software. Examples of SaaS include
Gmail, Microsoft Office 365, and Google Docs, which offer convenient and scalable solutions for email,
productivity, and collaboration without requiring extensive IT support.
3. Platform as a Service (PaaS): Platform as a Service (PaaS) provides a cloud-based platform that enables
developers to create, run, and manage applications without dealing with the complexities of infrastructure
management. PaaS solutions offer a complete development environment that includes tools, libraries, and
frameworks necessary for application development. This model allows businesses to reduce costs associated
with software licenses, hardware, and middleware, while also accelerating the development and deployment
of applications. Companies like Microsoft Azure, Amazon EC2, and Google App Engine exemplify PaaS by
offering robust platforms that support a wide range of programming languages and development tools.
 4. Infrastructure as a Service (IaaS): Infrastructure as a Service (IaaS) provides virtualized computing
resources over the Internet, offering businesses access to servers, storage, and networking capabilities on a
pay-as-you-go basis. This model enables organizations to scale their infrastructure according to demand
without investing in physical hardware. Clients can run and manage their applications while the cloud
provider handles the underlying infrastructure. IaaS solutions, such as Amazon Web Services (AWS), IBM
Cloud, and Microsoft Azure, deliver high scalability and flexibility, making it easier for businesses to adapt to
changing needs and reduce capital expenditures on IT infrastructure.
5. Analytics as a Service (AaaS): Analytics as a Service (AaaS) combines cloud computing with big data
analytics to offer a flexible and scalable platform for data analysis. AaaS provides tools and services that help
businesses extract insights from large volumes of data, supporting better decision-making and problem-
solving. This model covers all aspects of the analytics process, from data collection and storage to analysis
and visualization. By leveraging cloud-based analytics, businesses can access advanced analytical capabilities
without significant upfront investment in hardware and software. AaaS platforms enable data scientists and
analysts to work more efficiently, with examples including cloud-based analytics services that integrate
seamlessly with existing business processes.
Cloud computing supports business analytics by enabling scalable data warehouses and flexible, cost-effective
analytics services. Organizations can subscribe to these services on a pay-per-use basis, reducing the need for
significant upfront investments in infrastructure. This model is exemplified by companies like Amazon, Salesforce,
IBM Cloud, Microsoft Azure, and Google. The integration of cloud computing with business analytics allows for the
efficient handling of large data volumes and advanced analytical applications, facilitating better decision-making and
operational efficiency.

9. Cloud Deployment Models

Cloud deployment models refer to different approaches organizations can take to utilize cloud computing services
based on their specific needs and requirements. Here's a summary of the three common cloud deployment models:
Private Cloud: A private cloud is dedicated to a single organization, offering enhanced security and control over data
and applications. It operates within the organization's own data centers or through a dedicated third-party provider,
providing scalability and on-demand resource allocation but requiring internal management and higher costs.
Public Cloud: Public cloud services are managed by third-party providers and accessed over the internet by multiple
organizations. They offer cost-effective scalability, rapid deployment of resources, and reduced IT infrastructure
management overhead. However, organizations have less control over security and data privacy compared to private
clouds.
Hybrid Cloud: The hybrid cloud combines elements of both private and public clouds, allowing organizations to
optimize their IT infrastructure. It offers flexibility to allocate workloads between private and public environments
based on specific needs. Hybrid clouds require robust connectivity and compatibility between different cloud
platforms but provide versatility in balancing security, scalability, and cost efficiency.
10. Location-Based Analytics
Location-Based Analytics
Location-Based Analytics involves using data that includes geographical or spatial components to gain insights and
make informed decisions. This type of analytics leverages real-time or dynamic location data generated by devices
like smartphones, GPS, and sensors. By integrating location information with other business data, organizations can
optimize processes, improve customer targeting, enhance supply chain management, and gain a competitive edge.
Geospatial Analytics
Geospatial Analytics is a subset of location-based analytics that specifically focuses on analyzing static geospatial
data, such as maps, postal codes, and geographic boundaries. This data is often visualized using geographic
information systems (GIS) to identify patterns, trends, and relationships within a geographical context. Geospatial
analytics helps organizations in visualizing key performance indicators (KPIs), understanding regional market
dynamics, and making strategic decisions based on geographical factors.
Real-Time Location Intelligence
Real-time location intelligence leverages the constant stream of location data from devices such as cars, buses, mobile
phones, cameras, and navigation devices, captured through technologies like GPS, Wi-Fi, and cell tower triangulation.
This extensive collection of location-enabled data provides a massive database of historical and real-time information.
By analyzing these movement patterns, businesses can gain valuable insights into customer behaviors and make
informed decisions about promotions and pricing. Algorithms that reduce the dimensionality of location data uncover
trends and relationships, enabling intelligent predictions and effective customer targeting. In marketing, companies use
platforms to analyze mobile users' location trails, delivering targeted coupons and offers as customers pass by
retailers. Augmented reality applications, like Pokémon GO and the Candybar app, further enhance user engagement
by integrating virtual items with real-world locations, providing both a gaming experience and a powerful marketing
tool. This real-time location intelligence is pivotal for organizations, driving innovative applications and improved
decision-making in operational and marketing strategies.
Analytics Applications for Consumers
The growth of smartphone apps has created opportunities for developing consumer-focused applications that use
analytics to help users make better decisions. Unlike organizational apps that mine consumer data for marketing, these
apps directly benefit the consumer by employing specific analytics.
1. Navigation and Traffic:
o Waze: A social navigation app that uses user-generated data to alert drivers about accidents, police
checkpoints, and other road issues. Acquired by Google, Waze aggregates and shares this information
with users.
2. Reviews and Ratings:
o Yelp: Allows users to submit and view reviews and ratings for businesses and products, helping
consumers make informed choices based on aggregated social data.
3. Predictive Parking:
o ParkPGH: Developed in collaboration with Carnegie Mellon University, this app predicts parking
availability in Pittsburgh by updating space availability every 30 seconds and using historical data and
current events to forecast future availability. This concept has led to many similar parking apps in
major cities.
4. Productivity Enhancements:
o Gmail: Google’s email app uses analytics to generate automated response suggestions based on
billions of email transactions, helping users respond more efficiently.
11. An Illustrative Big Data Technology Platform: Teradata Vantage
Introduction
Teradata Vantage is a sophisticated analytics platform that integrates analytic engines and functions, allowing users to
implement analytics with preferred data science languages (e.g., SQL, Python, R) and tools (e.g., Teradata Studio,
Jupyter Notebook) on any volume or type of data. This platform can be used by various analytics professionals (e.g.,
Data Scientists, Business Analysts) in different environments (On-Premises, Private Cloud, Public Cloud).
Key Components of Vantage
Understanding Vantage involves five main concepts:
1. Analytics Engines and Functions:
o Analytic Engines: These are comprehensive frameworks with software components integrated into
containers (like Docker) to deliver advanced analytics. Components include advanced analytics
functions, data access points, integration with visualization and workflow tools, built-in management
tools, and a scalable environment.
o Three Analytic Engines:
▪ NewSQL Engine: Includes high-speed analytics functions such as nPath, Sessionization, and
various scoring functions.
▪ Machine Learning Engine: Offers over 120 prebuilt functions for tasks like sentiment
analysis and predictive failure analysis.
▪ Graph Engine: Provides functions for analyzing relationships within networks, useful for
social network analysis, fraud detection, and threat identification.
2. Data Storage and Access:
o Vantage includes a Teradata MPP Database and a high-speed data fabric (Teradata QueryGrid™ and
Presto™) that connects to external data sources (e.g., Oracle, Hadoop, Cassandra). It supports various
data formats including relational, spatial, temporal, XML, JSON, Avro, and time-series.

3. Analytic Languages and Tools:

o Vantage supports SQL, R, and Python, and can execute analytics through tools like Teradata Studio, R
Studio, and Jupyter Notebooks.
4. Deployment:
o Vantage provides consistent analytic processing across different deployment options, including
Teradata Cloud, public cloud, and on-premises installations.
5. Usage:
o Vantage is designed for diverse analytics users. Prebuilt functions make it accessible to citizen data
scientists and business analysts, while data scientists can use familiar tools like R Studio and Jupyter
Notebooks. Users can also create apps with Teradata AppCenter for visualizations.
Example Usage
A global retailer used Teradata Vantage's machine learning algorithms to improve search results on their website. This
led to a significant increase in revenue from online sales by accurately matching search results with customer needs,
resulting in over $1.3 million in additional revenue during a two-month holiday period.
Summary
Teradata Vantage is a versatile and powerful analytics platform designed to handle any data volume and type,
supporting multiple analytics personas across various deployment environments. It integrates advanced analytic
engines, diverse data storage capabilities, and popular data science tools, enabling efficient and scalable analytics
solutions.