Module-11

Data Analysis and Modelling

-----------------------------------------------------------------------------------------------------------------------------
11.1
Process, Benefits and Types of Data Analysis
-----------------------------------------------------------------------------------------------------------------------------
Data Analytics
Meaning  Data analytics involves examining raw datasets to uncover patterns
and draw meaningful insights.
 It helps businesses optimize operations, improve customer
understanding, and enhance DM.
Process of 1. Grouping Data
Data 2. Collecting Data
Analytics 3. Organizing Data
4. Cleaning Data
5. Choosing the Right Analytics Approach:
 Descriptive Analytics – Summarizes past data.
 Diagnostic Analytics – Identifies reasons behind outcomes.
 Predictive Analytics – Forecasts future trends.
 Prescriptive Analytics – Recommends actions based on insights.

11.2
Data Mining and Implementation of Data Mining
-----------------------------------------------------------------------------------------------------------------------------
Data Mining
Meaning  Data mining, also known as Knowledge Discovery in Data (KDD),
involves extracting patterns and valuable insights from large
datasets.
 With the rise of big data and AI, data mining helps businesses improve
decision-making, detect fraud, and identify trends.
Process of 1. Setting Business Objectives:
Data Mining 1. Define the problem and determine key data queries.
2. Data Preparation:
2. Gather, clean, and structure data by removing inconsistencies and
outliers.
3. Model Building & Pattern Mining
3. Use machine learning techniques like clustering, classification, and
regression to find insights.
4. Evaluation & Implementation
4. Validate findings and apply insights to achieve business goals.
Key Data Association Identifies relationships between data points (e.g.,
Mining Rules market basket analysis).
Techniques Neural Networks Mimics human brain processing for deep learning
applications.
Decision Trees Uses classification/regression to predict outcomes
through a tree-like structure.
K-Nearest Groups data points based on their proximity to similar
Neighbour (KNN) points

Page 1 of 8
Applications 1. Detecting Financial Crimes – Identifies fraudulent activities like
in Finance & money laundering.
Management 2. Loan & Credit Analysis – Predicts loan repayment likelihood and
assesses customer credit.
3. Target Marketing – Enhances customer segmentation and marketing
strategies.
4. Data Warehousing – Organizes vast transaction records for better
business insights.

Page 2 of 8
 11.3
Analytics and Model Building
(Descriptive, Diagnostic, Predictive, Prescriptive)
-----------------------------------------------------------------------------------------------------------------------------
Factor Descriptive Diagnostic
Predictive Prescriptive
Analytics Analytics
Analytics Analytics
Meaning Summarizes Analyzes
Uses past Provides
past data to data to determine
historical recommendations
understand what why something
data to on what actions
happened. happened (Why?)
predict should be taken.
future
outcomes.
Objective Provides Identifies causes Forecasts Suggests the best
insights into and relationships potential course of action to
past trends and in past data. future optimize future
patterns. trends and outcomes.
events.
Applicability 1. Reporting, 1. Root cause 1. Forecast, 1. DM
2. Trend analysis 2. Risk 2. Optimization,
analysis 2. Troubleshoting Assessm and automation.
3. Performanc 3. Identifying ent,
e reviews. correlations. 3. Strategic
planning.
Techniques 1. Data 1. Drill-down 1. Statistica 1. AI,
Used aggregation, analysis, l 2. optimization
2. Summarizat 2. Correlation modellin algorithms,
ion, analysis, and g, 3. Simulation
3. Visualizatio 3. Data mining. 2. ML, models,
n (charts, 3. Regress’ 4. Decision trees.
graphs, n
reports). 4. Algo.
Complexity Low Moderate High Very high
Examples 1. Monthly 1. Identifying 1. Forecasti 1. Recommending
sales reports why sales ng next the best pricing
2. Website dropped in a qtr. sales strategy for
traffic specific 2. Predict max. revenue
analysis quarter customer 2. Suggesting the
3. Social 2. Finding churn optimal
media reasons behind 3. Estimate inventory level
engagement a high DD for to min. waste
metrics employee products 3. AI-powered
attrition rate in personalized
3. Analysing different mkting.
customer seasons campaign
complaints to recommend’n
detect root
causes

Page 3 of 8
 11.4
Standards for Data Tagging and Reporting (XML, XBRL)
-----------------------------------------------------------------------------------------------------------------------------
Feature XML XBRL (Extensible Business
(Extensible Mark-up Language) Reporting Language)
Definition A mark-up language for storing, A specialized XML-based language
transferring, and structuring data in designed for financial and business
a human & machine-readable reporting.
format.
Purpose General-purpose data Standardizes financial data
representation and exchange.
reporting and analysis.
Structure Uses predefined tags for financial
Uses customizable tags to define
data structures. data to ensure consistency.
Validation Uses
Requires XML Schema (XSD) for predefined financial
structure validation. taxonomies for validation and
standardization.
Applications Web services, APIs, document Financial reporting, regulatory
formats (RSS, SVG, XHTML), and compliance, and business data
communication protocols (SOAP, exchange.
XMPP).
Automation Primarily used for manual and Fully automates financial
automated data exchange but reporting, analysis, and comparison.
requires additional processing.
Benefits Universal, flexible, and widely Reduces errors, improves efficiency,
adopted across industries. and facilitates financial decision-
making.
Examples RSS feeds, AJAX requests, Open ROC filings, financial statements,
Document formats. and investor reports.

Page 4 of 8
 11.5
Cloud Computing, Business Intelligence, Artificial Intelligence, Robotic Process
Automation and Machine Learning
-----------------------------------------------------------------------------------------------------------------------------
Cloud Computing [S-P-N]
Meaning  Cloud computing is the delivery of computing services (such as
Storage (S), Processing (P), And Networking (N) over the Internet
instead of relying on local servers.
Types of Private Cloud Exclusive to a single organization, offering high
Cloud security and control.
Computing Public Cloud Services and resources shared over the Internet,
providing scalability and cost-efficiency.
Hybrid Cloud Combines private and public cloud advantages,
allowing flexibility for sensitive and large-scale
computing tasks.

Business Intelligence (BI) (Analysis + Management of Data)

[Combination of Analytics + Data Mgt & Reporting tools + Methodologies]
Basic and  A technology-driven process + for analysing data + Delivering
Significance actionable information + Helps executives, managers and workers for
DM.
 As part of the BI process,
Data collection (IT internal & External)  Run queries against the
data  Create data visualizations,
 BI dashboards and reports to make the analytics results available
to business users for operational decision-making and strategic
planning
Main objective Drive to make DM
(Benefits)  ↑ Revenue
 ↑ Operational efficiency
 Gain competitive advantages over business rivals
 Spot business problems that need to be addressed;
 Identify emerging business and market trends;
 Develop stronger business strategies
BI Forms of Advanced analytics, such as
 Data mining (The process of sorting through large data sets to
identify patterns and relationships that can help solve business
problems through data analysis
 Predictive analytics (Use of statistics and modelling techniques to
make predictions about future outcomes and performance)
 Text mining (Analysing vast collections of textual materials to
capture key concepts, trends and hidden relationships)
 Statistical analysis
 Big data analytics (The process of collecting, examining, and
analysing large amounts of data to discover market trends, insights
etc)

Page 5 of 8
Artificial Intelligence (AI)
Meaning  Artificial refers to something which is made by humans or non-
natural things and Intelligence means the ability to understand or
think.
 AI is not a system but it is implemented in the system.
Types
Based on Capabilities of AI
Narrow AI  It is designed and trained on a specific task or a narrow range tasks
(Weak AI)  Voice assistants like Siri or Alexa that understand specific
commands.
 Facial recognition software used in security systems.
 Recommendation engines used by platforms like Netflix or Amazon.
General AI  AI systems that have human intelligence and abilities to perform
(Strong AI) various tasks.
 Systems have capability to understand, learn and apply across a
wide range of tasks that are similar to how a human can adapt to
various tasks.
 Ex. Robots
Super AI  Super AI surpasses intelligence of human in solving-problem,
creativity, and overall abilities.
 Super AI develops emotions, desires, need and beliefs of their own
Based on Functionality of AI
Reactive  Reactive machines are the most basic form of AI. They operate purely
Machine based on the present data and do not store any previous experiences or
learn from past actions.
 Ex. IBM’s Deep Blue, Google’s AlphaGo
Limited  Limited Memory AI can learn from past data to improve future
Memory in AI responses.
 Most modern AI applications fall under this category.
 Ex. Self-driving cars, Chatbots
Theory of  Theory of Mind AI aims to understand human emotions, beliefs,
Mind intentions, and desires
 Ex. Human-robot interaction, Collaborative robots
Self-  Self-Aware AI is an advanced stage of AI that possesses self-
Awareness AI consciousness and awareness
 Ex. Fully autonomous systems

Robotic Process Automation (RPA)

Meaning  RPA enables software robots ("bots") to automate repetitive, rules-
based tasks by mimicking human digital actions.
 Bots can copy-paste, extract data, perform calculations, log into
systems, interact with APIs, and more.
 Unlike humans, RPA bots work 24/7 with 100% accuracy and
reliability.
 Bots are easy to configure using simple commands like recording and
playback, making automation accessible without modifying existing
systems.

Page 6 of 8
 Key Differences among AI, ML, and Deep Learning
Feature AI ML DL
Definition  The broad  A subset of AI  A subset of ML
concept of that enables using neural
machines machines to learn networks to process
simulating from data. large data with
human minimal human
intelligence. intervention.
Goal  To create  To enable  To automate feature
intelligent computers to learn extraction and
machines that from patterns and improve learning
can think and act make decisions. with deep neural
like humans. networks.
Human  May require  Requires human  Minimal human
Intervention feature selection intervention; can
and structured learn from raw,
data. unstructured data.
Data  Can work with  Requires  Requires large
Requirement smaller datasets structured, amounts of data to
and predefined labelled data for train deep neural
rules. training. networks effectively.
Complexity  Less complex, Moderately Highly complex,
complex,
Example  Chatbots,  Spam filters,  Image recognition,
Applications  Expert systems, recommendation voice assistants,
 Self-driving cars systems,  Natural language
(overall  Fraud detection. processing (NLP).
intelligence).

Approaches to Machine Learning (ML)

Supervised  Uses labeled data with input-output pairs to train models.
Learning  Examples: Email spam filters, image classification, and price
prediction.
Unsupervised  Uses unlabeled data to find hidden patterns (clustering, grouping).
Learning  Example: Customer segmentation, anomaly detection, and
recommendation systems.
Semi-  Uses a mix of a small amount of labeled data and a large amount of
Supervised unlabelled data to improve accuracy.
Learning  Example: Medical diagnosis, fraud detection.
Reinforcement  Machines learn by interacting with an environment and receiving
Learning rewards or penalties.
 Example: Self-driving cars, AI playing games like chess.
Dimensionality  Reduces the number of variables while preserving key information.
Reduction  Example: Principal Component Analysis (PCA) simplifies complex
datasets for visualization and efficiency.

Page 7 of 8
 11.6
Model vs. Data-driven Decision-making
-----------------------------------------------------------------------------------------------------------------------------
Artificial Intelligence (AI), there are two main approaches:
Data-Driven  Focuses on improving Data Quality, Governance, and
Approach Management to enhance AI performance.
Model-Driven  Focuses on developing better algorithms and models to improve
Approach AI capabilities.
 It is advanced significantly as compared to Data-driven approach
Real Word AI applications during COVID-19:
Example  AI was used to detect COVID-19 from medical images, predict
disease progression, and estimate fatalities.
 However, many studies lacked high-quality data, proper
validation, and demographic diversity, limiting real-world
application.
Key Points  No matter how advanced an AI model is, poor data will always
lead to poor performance.
 For AI to be truly effective, Systematic Data Collection,
Management, And Validation are crucial.

Page 8 of 8