0% found this document useful (0 votes)
2 views39 pages

BIDA Module 2

The document provides an overview of data analytics, including the nature of data, the analytics process, and the importance of data preprocessing. It discusses big data, its characteristics (7 V's), and technologies used for processing and analyzing large datasets. Additionally, it covers statistical modeling, regression techniques, and time series forecasting as essential components of business analytics.

Uploaded by

meghanapa3
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
2 views39 pages

BIDA Module 2

The document provides an overview of data analytics, including the nature of data, the analytics process, and the importance of data preprocessing. It discusses big data, its characteristics (7 V's), and technologies used for processing and analyzing large datasets. Additionally, it covers statistical modeling, regression techniques, and time series forecasting as essential components of business analytics.

Uploaded by

meghanapa3
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

DESCRIPTIVE

ANALYTICS I
THE NATURE OF DATA IN
ANALYTICS
Data: Raw facts and figures without
context. When processed or analyzed, data
becomes information.
In analytics, data is the foundation upon
which insights, predictions, and decisions
are built.
DATA IN THE ANALYTICS
PROCESS
Data Collection
Data Preparation (Cleaning &
Transformation)
Data Exploration and Visualization
Modeling and Analysis
Interpretation and Communication of
Results
A SIMPLE TAXONOMY OF
DATA
A taxonomy of data is a structured
way to classify data based on its
characteristics, structure, and usage.
It helps in understanding how
different types of data are handled
in analytics and data science.
THE ART AND SCIENCE OF DATA
PREPROCESSING
Data preprocessing is the critical first step
in the data analytics and machine learning
pipeline. It involves transforming raw data
into a clean, structured, and usable
format.
It’s often referred to as both an art and a
science because it combines technical
rigor with problem-solving creativity.
DATA PREPROCESSING STEPS

Data
Raw Data Data Cleaning
Consolidation

Data Well Formed


Data Reduction
Transformation Data
BIG DATA
Big Data refers to extremely large
and complex datasets that are
difficult to process, store, manage,
or analyze using traditional data
processing tools.
FUNDAMENTALS OF BIG DATA ANALYTICS
Big Data Analytics:

It is the process of collecting, organizing,


and analyzing large volumes of data (big
data) to discover patterns and insights that
support better decisions and strategic
business moves.
FUNDAMENTALS OF BIG DATA ANALYTICS
Type Question Focus Outcome
Answered

Descriptive What happened? Past Insight and


reporting

Diagnostic Why did it Causes of past Root cause


happen? events identification

Predictive What will Future Forecasts and


happen? probabilities

Recommended
Prescriptive What should we Decision- actions and
do? making strategy
7 V’S
OF
BIG DATA
7 V’S OF BIG DATA
Volume
Refers to the amount of data being generated.
Big Data often deals with terabytes, petabytes, or even zettabytes of
information.
Sources: social media, IoT devices, business transactions, online
clicks, sensors.
Example: YouTube uploads more than 500 hours of video every
minute.

Velocity
Refers to the speed of data generation and processing.
In many cases, data must be analyzed in real time or near real time.
Example: Credit card fraud detection systems must process
transactions instantly to block suspicious activity.
7 V’S OF BIG DATA
Variety
Refers to the different types of data collected
The insights may come without structure. The total asset may
include many data types, from XML to video to SMS.
Example: An e-commerce company analyzing customer reviews
(text), purchase history (structured), and browsing behavior
(clickstream logs).

Veracity
Refers to the quality, reliability, and accuracy of data.
Big Data often contains inconsistencies, duplicates, and noise.
Example: Fake news or spam reviews on Amazon can distort
customer sentiment analysis.
7 V’S OF BIG DATA
Value
Refers to the usefulness and business impact of data.
Collecting big data is meaningless unless it creates insights,
decisions, or profits.
Example: Netflix uses viewing data to recommend movies,
reducing churn and increasing engagement

Variability
Refers to the changing nature of data meaning and flow.
Data may have different patterns depending on time, season, or
context.
Example: The keyword “apple” may mean fruit in one dataset and
technology brand in another.
7 V’S OF BIG DATA
Visualization
Refers to the presentation of big data insights in an
understandable way.
Visualization helps decision-makers grasp patterns,
trends, and anomalies quickly.
Tools: Tableau, Power BI, [Link], Python (matplotlib,
seaborn).
Example: COVID-19 dashboards that visualize global
cases, recoveries, and vaccination rates in real time.
BIG DATA TECHNOLOGIES
Category Technologies/Tools

Storage HDFS, NoSQL, Data Lakes

Processing Hadoop, Spark, Flink, Storm

Integration Kafka, NiFi, Sqoop

Analysis R, Python, MLlib, Mahout

Visualization Tableau, Power BI, Qlik

Cloud Platforms AWS, Azure, GCP


BIG DATA AND STREAM ANALYTICS
Big Data refers to extremely large datasets that
cannot be easily managed, processed, or analyzed
using traditional tools.

Stream Analytics is the process of real-time


processing and analysis of continuous data
streams. Instead of waiting to collect and store
data in a database or data warehouse, it is
analyzed as it flows.
APPLICATIONS OF STREAM ANALYSIS
Application Area Stream Analytics Use Case Benefit

Banking & Finance Fraud detection Prevent loss and fraud

Personalize customer
Retail Real-time recommendations
experience

Manufacturing Predictive maintenance Reduce downtime

Traffic and resource


Smart Cities Improve urban efficiency
optimization
Early diagnosis, emergency
Healthcare Live health monitoring
alerts

Social Media Trend & sentiment analysis Real-time brand management

Faster and more efficient


Logistics Shipment and route tracking
delivery
DIFFERENCE BETWEEN BIG DATA
AND STREAM ANALYTICS
Feature Big Data Analytics Stream Analytics

Data Type Historical data (stored) Real-time/live data

Latency (Speed of
Insights) Minutes to hours Milliseconds to seconds

Processing Batch mode Continuous/real-time

Trend analysis, Fraud detection, live


Use Case historical reporting monitoring

Apache Kafka, Spark


Hadoop, Spark (batch),
Tools Hive Streaming, Flink, Azure
Stream Analytics
STATISTICAL MODELING FOR BUSINESS
ANALYTICS
Statistical modeling in business analytics is the process
of using mathematical models based on statistics to
analyze data, understand relationships, and make
predictions or decisions in a business context.

It is the use of quantitative methods and statistical


techniques to analyze data patterns, test hypotheses,
and support business decisions.
Model Type Key Function Common Use Case

Market segmentation, sales


Descriptive Describe data
summaries
Churn prediction, risk
Predictive Predict future outcomes
assessment
Pricing optimization, resource
Prescriptive Recommend actions
planning

Diagnostic Explain past events Root cause of revenue decline

Customer preference
Inferential Generalize from samples
estimation

Time Series Forecast over time Sales/demand forecasting


DESCRIPTIVE STATISTICS FOR DESCRIPTIVE
ANALYTICS

Descriptive Statistics is a foundational component


of Descriptive Analytics, which focuses on
summarizing and understanding historical data to
identify patterns, trends, and insights that inform
business decisions.
MEASURES OF CENTRAL TENDENCY

Measure Description Example Use

Sum of values divided


Mean (Average) Average monthly sales
by count

Middle value in sorted Middle income of


Median
data customers

Most purchased
Mode Most frequent value
product
MEASURES OF DISPERSION (SPREAD)

Measure Description

Range Difference between max and min values

Variance Variance measures how far each data point is from the mean,
Average of the squared differences from the mean

Standard Deviation Square root of variance, it shows how much data values deviate
(vary) from the mean (average)
SHAPE AND DISTRIBUTION
Measure Description Example Use

Measure of symmetry of a
Are sales data skewed to one
Skewness data distribution (mean =
side?
median = mode)

Measure of peakedness or Identify outliers in


Kurtosis
flatness performance

Number of customers by age


Frequency Distribution How often values occur
group

Analyze sales across product


Histograms/Boxplots Visual representations
categories
REGRESSION MODELING FOR INFERENTIAL
STATISTICS
Regression modeling is a powerful statistical tool used in
inferential statistics to understand relationships between
variables and to make predictions.

In the context of business analytics, regression helps draw


conclusions about the population from sample data and
supports data-driven decision-making.
INFERENTIAL STATISTICS
Inferential statistics involves using a sample of data to make
inferences or generalizations about a larger population.

Purpose:
To test hypotheses
To estimate relationships
To predict outcomes
REGRESSION MODELING
Regression modeling estimates the relationship between a dependent
variable and one or more independent variables. It helps determine
how changes in the inputs affect the output.

Uses:
Understand cause-effect relationships
Predict future values
Quantify impact of variables
TYPES OF REGRESSION MODELS IN
INFERENTIAL STATISTICS

Simple Linear Regression


Multiple Linear Regression
Logistic Regression
Polynomial Regression
Ridge Regression
Lasso Regression
TYPES OF REGRESSION MODELS IN INFERENTIAL
STATISTICS
Simple Linear Regression:
Finds a straight-line relationship between one independent variable (X)
and one dependent variable (Y)
Example: Predicting sales (Y) based on advertising spend (X)

Multiple Linear Regression


Extends simple linear regression to include two or more independent
variables (X1, X2, …)
Example: Predicting house price (Y) using size, location, and number of
bedrooms (X1, X2, X3)
TYPES OF REGRESSION MODELS IN INFERENTIAL
STATISTICS
Logistic Regression
Used when the outcome is categorical (Yes/No, True/False, 0/1). It predicts the
probability of an event happening.
Example: Predicting whether a customer will buy a product (Yes/No) based on
income and age.

Polynomial Regression
Fits a curved line (instead of a straight line) when the relationship between X
and Y is not linear.
Example: Predicting the growth of plants over time, where growth first
speeds up, then slows down.
TYPES OF REGRESSION MODELS IN INFERENTIAL
STATISTICS
Ridge Regression
Ridge regression is an improved version of linear regression used when there are many
predictors or when predictors are highly correlated (multicollinearity).
A type of regression that reduces overfitting by adding a penalty (shrinkage) on large
coefficients.
Example: Predicting sales using many factors (like price, ads, seasons, competitors), but
avoiding giving too much weight to any one variable.

Lasso Regression
Similar to Ridge but it can shrink some coefficients to zero, effectively selecting only the most
important predictors.
Example: Predicting house prices from 50 variables (Location, size, bedrooms, bathrooms,
garden, distance to schools, flooring type, wall paint color, etc.) but automatically removing
unimportant ones like “color of mailbox.”
LINEAR REGRESSION MODELING
1 Define the Problem

Collect and Prepare the Data 2

3 Visualize the Data


Fit the Linear
Regression Model 4

5 Evaluate the Model


Use the Model for
Prediction 6
Interpret and
7
Communicate Results
IMPORTANT ASSUMPTIONS IN LINEAR
REGRESSION
Linearity
Independence of Errors
Error (Residual)=Actual Value (Y)−Predicted
Value (Yˆ))
Constant Variance
Normality of Errors
No Multicollinearity
TIME SERIES FORECASTING
Time series forecasting involves using
historical data points measured over time
(usually in equal intervals) to predict future
values.
A time series is a sequence of data points
collected or recorded at successive time
intervals.
It is widely used in business, finance, weather,
and operations planning.
TIME SERIES FORECASTING TECHNIQUES

Naïve Method
Moving Average
Exponential Smoothing
Holt-Winters
Facebook Prophet
Machine Learning
TIME SERIES FORECASTING TECHNIQUES
Method Best For
Naïve Quick guess, baseline
Moving Average Smoothing data
Exponential Smoothing Recent changes
Holt-Winters Trends + seasonality

Machine Learning Complex or big data

Facebook Prophet Business data with patterns


THANK
YOU

You might also like