Module 1

Introduction to Business Analytics

Definition & Scope of business analytics
• Business analytics is the process of gathering, processing, and analyzing
data to support business decision-making.
• By utilizing techniques from statistics, data mining, and machine learning,
it identifies patterns, trends, and insights from both historical and real-
time data.
• These insights help businesses enhance operations, predict future trends,
and drive performance improvements.
• Business analytics is categorized into three types: descriptive analytics
(understanding past events/what has happened)
• predictive analytics (anticipating future outcomes/what is likely to happen)
• and prescriptive analytics (recommending actions/what should be done).
• It often employs tools like dashboards, data visualization, and modeling to
enable data-driven decision-making.
• Data Mining
• Data mining is the process of discovering patterns, relationships, and useful information from
large datasets using statistical, machine learning, and artificial intelligence techniques. It helps
in identifying hidden trends and insights that can be used for decision-making and prediction.
• Key Steps in Data Mining:
1.Data Collection – Gathering relevant data from different sources.
2.Data Cleaning – Removing errors, inconsistencies, and missing values.
3.Data Transformation – Converting raw data into a suitable format.
4.Pattern Discovery – Using algorithms to identify trends and relationships.
5.Evaluation & Interpretation – Analyzing and interpreting the findings.
• Machine Learning (ML)
• Machine Learning is a branch of artificial intelligence (AI) that enables computers to learn
from data and make decisions or predictions without being explicitly programmed. It involves
algorithms that identify patterns in data and improve their performance over time.
• Machine Learning is a type of AI that helps computers learn from data and make decisions without being directly programmed. It
uses patterns in data to improve over time, just like how humans learn from experience

• Types of Machine Learning:

1.Supervised Learning – The model is trained on labeled data (e.g., spam detection, loan
approval).
2.Unsupervised Learning – The model identifies patterns in unlabeled data (e.g., customer
segmentation, anomaly detection).
3.Reinforcement Learning – The model learns by interacting with an environment and
receiving feedback (e.g., robotics, game AI, self-driving cars).
• Key Applications of Machine Learning:
• Healthcare – Disease diagnosis, medical image analysis.
• Finance – Fraud detection, stock price prediction.
• Retail – Recommendation systems, demand forecasting.
• Manufacturing – Predictive maintenance, quality control.
• Autonomous Systems – Self-driving cars, smart assistants (e.g., Siri,
Alexa).
Analysis vs analytics

There is a technical difference between both

There is a subtle difference between the terms and

meaning of the words Analysis and Analytics.

Analysis - detailed examination of the elements or

structure of something

Ex: chemical Analysis- In a laboratory, a chemist conducts an analysis

of a water sample to identify the presence and concentration of
various chemicals or pollutants.
Analytics - the systematic computational analysis of
data or statistics

A company uses sales analytics to track and predict sales performance.

By analysing historical sales data; the company can forecast future sales,
optimize inventory levels, and tailor marketing strategies to improve sales
performance

Analysis can be done without numbers or data, such as

business analysis, psycho analysis, etc.

Whereas Analytics, even when used without the prefix

"Data", almost invariably implies use of data for
performing numerical manipulation and inference.
Business
• The aim of business analytics is data and
Analytics reporting—examining past business
performance and forecasting future business
performance.
vs.
• On the other hand, the business analysis
focuses on functions and processes- focusses
Business on improving business processes and solving
Analysis problems.
Importance of business analytics in
organizations
Informed Decision Making: BA provides insights into data
trends and patterns, enabling organizations to make better-
informed decisions. By analyzing past data, companies can
predict future outcomes and optimize strategies.

Improved Efficiency and Operations: Through process

optimization and performance measurement, BA helps identify
inefficiencies and bottlenecks in operations, leading to reduced
costs and improved workflow.

Competitive Advantage: Organizations that use business

analytics can identify emerging market trends, customer
preferences, and potential opportunities faster than
competitors, giving them an edge in the marketplace.

Customer Insights: By analyzing customer behavior,

preferences, and feedback, BA enables businesses to create
personalized experiences, improve customer satisfaction, and
increase loyalty.
• Financial Management: BA helps organizations track financial metrics,
improve budgeting, and forecast future financial performance, leading to
better resource allocation and financial stability.
• Risk Management: Through predictive analytics, organizations can
assess potential risks, identify vulnerabilities, and take preemptive
actions to mitigate them.
• Strategic Planning: With the help of data analytics, businesses can
better understand industry trends, competitor actions, and market
dynamics, aiding in long-term strategic planning and decision-making.
• Innovation and Growth: By uncovering new insights about products,
services, and market needs, business analytics fosters innovation and
helps organizations expand into new areas or improve their offerings
Types of business
analytics
 Descriptive
Analytics

• Refers to the process of analyzing historical data to understand what has

happened in the past within an organization or business.

• It provides insights into past performance, trends, and patterns, helping

decision-makers gain a better understanding of the business environment.

• Descriptive analytics doesn't predict future outcomes, but it provides a solid

foundation for decision-making by summarizing and interpreting historical
data.
 Key Aspects of
Descriptive
Analytics
• Data Summary: Descriptive analytics typically focuses on summarizing
large datasets into a more understandable and accessible format.

• Data Exploration: It involves exploring and visualizing data to identify

key trends, patterns, and relationships that have occurred over time.

• Reporting: This form of analytics often results in reports, dashboards,

and data visualizations that showcase important metrics and trends
 1. Data Aggregation:
• The process of combining data from
various sources to provide a holistic
view. This can include summarizing
data into metrics like averages, totals,

Tools and and percentages.

• Tools: Excel, Google Sheets, SQL
2.databases.
Data Visualization:
Techniques • Visual representation of data through
charts, graphs, and dashboards to

for make it easier to identify patterns and

trends.
• Tools: Tableau, Power BI, QlikView,

Descriptive Google Data Studio, Excel.

3. Statistical Analysis:

Analytics • 1.Using statistical methods to

calculate key metrics such as
averages, standard deviations,
frequencies, and correlations.
• 2. Tools: R, Python (with libraries like
Pandas, NumPy, Matplotlib), SPSS,
SAS.
 4.Descriptive Statistics:
• These are basic statistics used to summarize data, such as:
• Mean (average)

Tools and •
•
Median (middle value)
Mode (most frequent value)

Technique •
•
Standard Deviation (measure of variation)
Frequency Distribution (how often certain values occur)
s for • Tools: Excel, R, Python (with Pandas, SciPy).

Descriptiv 5. Reporting and Dashboards: :

• Creating static or interactive reports and dashboards to
e present data summaries and key performance indicators (KPIs)
to stakeholders.
Analytics • Tools:Tableau, Power BI, Google Data Studio, Excel, Zoho
6.Analytics.
Pivot Tables:
• A powerful feature used for summarizing, analyzing, exploring,
and presenting data, especially in business contexts.
• Tools:Excel, Google Sheets.
 7. SQL Queries::

• Writing queries to extract, aggregate, and

analyze data directly from databases.
• Tools:SQL Server, MySQL, PostgreSQL,
Tools and Oracle Database.

Technique 8. Time Series Analysis:

s for
• Analyzing data that is collected over time to
identify trends and patterns, often used in
business for tracking performance over
Descriptiv specific periods.
• Tools: Excel, Python (with libraries like

e
Pandas, Stats models), R.
9. Text Mining:

Analytics • For analyzing unstructured data, such as

customer reviews or social media comments,
to identify patterns or trends.
• Tools: Python (with NLTK, SpaCy), R (with tm
package).
 1. Trend Analysis: Analyzing data points over a period to
identify consistent patterns or trends

(e.g., sales growth, market share, or customer behavior).

2. Customer Segmentation: Categorizing customers into

different segments based on common characteristics or
Methods for behaviors to identify trends in each segment.

Descriptive 3. Comparative Analysis: Comparing performance metrics

Analytics across different time periods, geographical locations, or product

categories to spot areas of improvement or success.

4. Correlation Analysis: Analyzing relationships between two

or more variables to understand how they move together

(e.g., the relationship between marketing spend and sales).

5. Frequency Distribution: Examining the frequency of

various outcomes (such as how often customers make
purchases or how often a particular product is sold) to identify
patterns.
 Diagnostic Analytics

• Is a type of data analysis that focuses on

understanding the reasons behind past
outcomes.
• While descriptive analytics answers "what
happened”
• diagnostic analytics aims to answer, "why
did it happen?"
• It helps organizations identify the root
causes of events, trends, or anomalies
that occurred in the past, providing
deeper insights into underlying factors
influencing performance.
 Cause-and-Effect Analysis:
Diagnostic analytics goes beyond
identifying patterns; it seeks to find the
reasons or causes behind those
Key Aspects patterns or outcomes.
of Problem Solving: This type of
analytics helps identify areas of
Diagnostic concern or unexpected results and
provides actionable insights for
Analytics improving processes or performance.
Data Exploration: It requires
examining data from various
perspectives to identify correlations,
trends, and factors that may have
contributed to a specific outcome.
Tools and Techniques for Diagnostic
Analytics
Root Cause Analysis (RCA):
A method of identifying the
• Tools: Fishbone diagram (Ishikawa
underlying causes of a problem or diagram), 5 Whys technique, Failure Mode
event. It often involves asking "Why?"
multiple times to dig deeper into the and Effects Analysis (FMEA).
factors contributing to an issue.
Correlation Analysis:
Analyzing relationships between
variables to determine how one • Tools: Excel, R (with the cor() function),
variable might influence or correlate
with another. This can help identify
Python (with Pandas and NumPy).
potential causes of a particular
outcome.
Data Mining:
Using advanced techniques to • Tools: Python (with libraries like Scikit-learn,
discover patterns, correlations, and
relationships in large datasets, TensorFlow), R, SAS.
which may help explain why certain
trends or outcomes occurred.
Tools and Techniques for Diagnostic
Analytics
Regression Analysis:
A statistical technique used to
• Tools: R, Python (with Scikit-
understand relationships between
dependent and independent variables. It
can help identify the strength and
learn, Statsmodels), SPSS,
nature of relationships, helping
diagnose factors influencing a given
outcome.:
Excel.
Anomaly Detection:
This technique involves identifying
data points that deviate significantly • Tools: Python (with Scikit-
from normal patterns. It can help
identify unusual occurrences or
outliers, which can then be explored to
learn, Keras), R, TensorFlow.
understand their causes.
Trend Analysis:
By analyzing trends in data over
time, diagnostic analytics can
• Tools: Excel, Python (with
uncover patterns that may indicate
the reasons for performance
Pandas, Matplotlib), R.
fluctuations.
Tools and Techniques
Time Series
Decomposition: for Diagnostic
• Tools: R (with forecast package),
Analytics
Analyzing time-series data to
decompose it into trend, seasonality,
Python (with Statsmodels).
and residual components. This method
helps in identifying what factors are
influencing changes in data over time.
Segmentation Analysis:
Dividing data into different groups or • Tools: Python (with Pandas, Scikit-
segments based on certain
characteristics or behaviors to
understand why certain segments
learn), R, Excel.
behave differently.
Hypothesis Testing:
Using statistical methods to test
hypotheses and determine whether
• Tools: R, Python (with SciPy,
observed outcomes are due to
chance or if there are significant
Statsmodels), SPSS.
underlying causes.
Data Visualization:
Visual representations of data
through charts and graphs that
• Tools: Tableau, Power BI, Qlik,
highlight relationships, trends,
and potential causes behind Python (with Matplotlib, Seaborn)
data anomalies or trends
 Methods for
Diagnostic
Analytics
1. Cause-and-Effect Diagrams: Tools like fishbone diagrams (Ishikawa) and flowcharts help in
visually breaking down complex problems to uncover the root causes.

2. Pareto Analysis: The 80/20 rule, where 80% of problems come from 20% of causes. This
method is used to identify which issues are the most critical and worth addressing.

3. Drill-Down Analysis: This technique involves drilling down into data to identify subcomponents
and details that may reveal causes behind a high-level outcome or trend.

4. Multivariate Analysis: Using multiple variables simultaneously to understand how different

factors may interact and influence a particular outcome.

5. Benchmarking: Comparing current performance against industry standards or past

performance to identify gaps and understand the causes behind deviations.

6. Scenario Analysis: Using historical data to analyze different "what-if" scenarios to understand
how different actions or events might impact outcomes.
Predictive Analytics
Predictive Analytics is a branch of advanced analytics that
uses historical data, statistical algorithms, machine learning
techniques, and artificial intelligence (AI) to predict future
events or outcomes.
Unlike descriptive or diagnostic analytics, which focus on
understanding past events, predictive analytics forecasts
what is likely to happen in the future based on patterns and
trends identified in historical data.
It is widely used in various industries to anticipate risks,
opportunities, and behaviors.
Key Aspects of Predictive Analytics:
Forecasting Future Events: Predictive analytics is
primarily concerned with forecasting outcomes based on
historical and current data.

Probability and Risk: It often involves estimating the

likelihood of different future scenarios, helping
organizations understand the probability of various
outcomes.

Data-Driven Decision Making: Predictive analytics

uses data to inform decisions, enabling organizations to
take proactive actions to mitigate risks or seize
opportunities
 Tools and Techniques for Predictive Analytics

Regression Analysis: Machine Learning:

A statistical method for In predictive analytics, Time Series Analysis:
modeling the relationship machine learning (ML)
algorithms are used to build A statistical method used to
between a dependent variable analyze time-ordered data to
(what you are trying to predict) models that learn from data
and improve over time. Popular forecast future values based
and one or more independent on past patterns. Time series
variables (predictors). Linear algorithms include decision
trees, random forests, support models account for trends,
and logistic regression are seasonality, and cycles in the
common forms of regression vector machines (SVM), and
neural networks. data.
analysis used for prediction.

Tools: Python
(with Scikit-learn,
Tools: R, Python Tools: R (with the
TensorFlow,
(with Scikit-learn, forecast package),
Keras), R, SAS,
Statsmodels), Python (with
IBM Watson,
SPSS, Excel. Statsmodels,
H2O.ai.
Prophet), Excel.
Tools and Techniques for
Predictive Analytics

Neural Networks:
A class of machine learning models inspired Decision Trees and Random Forests: Support Vector Machines (SVM):
Clustering:
by the human brain that are particularly Decision trees break down a dataset into A supervised learning algorithm that can be
While not strictly predictive, clustering
useful for handling large datasets with smaller subsets based on certain criteria, used for both classification and regression
techniques (like K-means or DBSCAN) group
complex patterns. Deep learning helping to predict outcomes. Random tasks. SVM works by finding the hyperplane
data into similar clusters that can help
techniques, such as Convolutional Neural forests are an ensemble method that builds that best separates the data into
make predictions by identifying patterns
Networks (CNNs) and Recurrent Neural multiple decision trees and averages their categories, making it useful for binary
within each cluster.
Networks (RNNs), are used for more predictions for greater accuracy. classification problems.
advanced predictions.

Tools: Python
Tools: Python
(with Tools: Python Tools: Python
(with Scikit-
TensorFlow, (with Scikit- (with Scikit-
learn, K-
Keras, learn), R, SAS. learn), R.
means), R.
PyTorch), R.
Tools and Techniques for Predictive Analytics

Natural Language Processing

(NLP):
Techniques from NLP, such as • Tools: Python (with NLTK,
sentiment analysis, can be used to
predict customer behavior, stock
market trends, and more, based on
SpaCy), R, IBM Watson.
unstructured textual data.

Ensemble Methods:
These methods combine multiple
• Tools: Python (with Scikit-
predictive models to improve accuracy.
Common ensemble techniques include
boosting (e.g., Gradient Boosting
learn), R, XGBoost,
Machines) and bagging (e.g., Random
Forest). LightGBM.
Monte Carlo Simulation:
A method used to understand the • Tools: Excel, Python (with
impact of risk and uncertainty in
prediction models by simulating
multiple scenarios and outcomes.
NumPy, SimPy), MATLAB
Methods for Predictive Analytics
Data Preprocessing: Before applying predictive models, data must be cleaned and
transformed. This step includes handling missing values, normalizing data, encoding
categorical variables, and removing outliers.

Feature Selection: Identifying the most relevant features (variables) for making predictions.
Feature selection helps reduce the complexity of the model and improves its accuracy.

Model Training: This involves splitting the data into training and testing sets to build and test
the model. The training set is used to train the model, while the testing set evaluates its
performance.

Model Evaluation: After training the model, its performance must be evaluated. Common
metrics for regression models include Mean Absolute Error (MAE), Mean Squared Error (MSE),
and R-squared. For classification tasks, metrics such as accuracy, precision, recall, F1-score,
and ROC-AUC are used.

Hyperparameter Tuning: The performance of a predictive model can be improved by

optimizing the hyperparameters (settings that control the learning process). Techniques like
grid search or random search are used for this purpose.

Real-Time Prediction: Once a predictive model is trained, it can be deployed for real-time
predictions on new incoming data. This is commonly used in industries like e-commerce,
finance, and healthcare.
Prescriptiv
e analytics
1. Is a type of data analytics that focuses on providing actionable
recommendations and determining the best course of action for
specific scenarios.

2. It goes beyond descriptive and predictive analytics by not only

understanding past events or predicting future trends but also
suggesting the optimal decisions to achieve desired outcomes.
 Key Features of Prescriptive Analytics

1. Decision-Making Guidance: It provides recommendations for

actions based on the analysis of data.
2. Optimization: Often includes finding the most efficient way to
achieve goals, such as maximizing profit, minimizing costs, or
improving customer satisfaction.
3. Scenario Modeling: Helps decision-makers evaluate various
"what-if" scenarios and choose the best one.
4. Proactive Solutions: Enables businesses to act before
problems occur or opportunities are missed.
 IBM Decision Optimization

To • Features: Provides tools for mathematical optimization and prescriptive modeling.

• Use Cases: Supply chain optimization, scheduling, and resource allocation.

ols
• Methods: Uses linear programming, mixed-integer programming, and constraint programming.

for Google OR-Tools

• Features: An open-source optimization tool developed by Google.

Pr • Use Cases: Transportation logistics, task scheduling, and route optimization.
• Methods: Includes solvers for linear programming, integer programming, and constraint
es satisfaction problems.

cri
Microsoft Azure Machine Learning

• Features: Includes prescriptive analytics capabilities for businesses using cloud-based services.
pti • Use Cases: Demand forecasting, dynamic pricing, and personalized marketing.
• Methods: Combines predictive modeling with optimization techniques.

ve SAP Analytics Cloud

An • Features: A comprehensive tool offering prescriptive capabilities within its analytics suite.

aly • Use Cases: Financial planning, sales optimization, and workforce management.
• Methods: Scenario planning and optimization integrated with business intelligence.

tic Tableau with R or Python Integration

s • Features: Combines visualization with prescriptive modeling when integrated with programming
languages like R and Python.
• Use Cases: Decision support systems in marketing, healthcare, and operations.
• Methods: Data-driven recommendations using machine learning and optimization algorithms.
Techniques and Methods for Prescriptive Analytics
1. Optimization Models

• What It Does: Determines the best possible outcome within a set of constraints.
• Methods: Linear programming, integer programming, and goal programming.
• Applications: Supply chain management, production scheduling, and budget allocation.

2. Simulation Models

• Model'st Does: Models complex systems to simulate various scenarios and predict the impact of different
actions.
• Methods: Monte Carlo simulation and agent-based simulation.
• Applications: Risk assessment, portfolio management, and capacity planning.
3. Decision Trees

• What It Does: Visual representation of decisions and their potential outcomes.

• Methods: CART (Classification and Regression Trees) and ensemble methods.
• Applications: Strategic planning, loan approval, and medical diagnosis.

4. Heuristic and Rule-Based Systems

• What It Does: Uses predefined rules or heuristics for decision-making in situations with limited data or
time.
• Methods: Expert systems and fuzzy logic.
• Applications: Customer service chatbots and fraud detection.
 Techniques and Methods for Prescriptive Analytics

5. Machine What It Does: Incorporates predictive models with prescriptive suggestions to automate
decision-making.
Learning Methods: Reinforcement learning, neural networks, and Bayesian networks.
Algorithms Applications: Dynamic pricing, recommendation engines, and autonomous vehicles.

What It Does: Examines multiple scenarios to identify the best course of action under
6. Scenario uncertainty.

Analysis Methods: "What-if" analysis and stress testing.

Applications: Financial forecasting and disaster recovery planning.

7. Natural What It Does: Extracts actionable insights from unstructured text data.
Language Methods: Text summarization, sentiment analysis, and intent recognition.
Processing Applications: Legal case analysis, contract reviews, and automated customer
responses.
(NLP)
Real-World Examples of
Prescriptive Analytics
Scenario: A retailer uses prescriptive analytics to optimize inventory
levels.
1. Retail: Action: By combining demand forecasting (predictive) with optimization
algorithms, the system recommends stock adjustments and reorder
points to minimize stockouts and reduce holding costs.
Scenario: Hospitals use prescriptive analytics to allocate resources
2. during peak times.
Healthcare: Action: Simulation models predict patient inflows, and optimization tools
recommend nurse staffing levels to ensure optimal patient care.
Scenario: Logistics companies optimize delivery routes.
3.
Transportat Action: Tools like Google OR-Tools calculate the shortest and fastest
routes, considering traffic, weather, and delivery deadlines, to reduce fuel
ion: costs and improve efficiency.
Scenario: Utility companies optimize power grid operations.
4. Energy: Action: Predictive models forecast energy demand, and prescriptive
models determine the best mix of renewable and non-renewable energy
sources to meet demand cost-effectively.

Scenario: Investment firms manage portfolio risks.

5. Finance: Action: Prescriptive analytics models recommend asset allocations based
on predicted market movements and risk tolerance levels.