Rangi Liyanage

[email protected]
MSc in Information Systems- SLIIT
BBA (Hons) Specialize in Business Analytics- SLIIT
 Introduction to the module

 Business, Data and Analytics

 Big Data

 What is Accounting Analytics

 Data Visualization
 What is a model?

 Types Of Decision Models

 Regression Analysis

 Machine Learning Concepts

 Optimization
 An abstraction or representation of a real system, idea, or object

 Captures the most important features

 Can be a written or verbal description, a visual display, a mathematical formula, or

a spreadsheet representation

 Used to understand, analyze, or facilitate decision making.

 Descriptive analytics
- uses data to understand past and present

 Predictive analytics
- predicts the future

 Prescriptive analytics
- uses optimization techniques
Types of
Business
Analytics
 Retail Markdown Decisions
 Most department stores clear seasonal inventory by reducing prices.

 The question is:

 When to reduce the price and by how much?

 Descriptive analytics: examine historical data for similar products (prices, units sold,
advertising, …)
 Predictive analytics: predict sales based on price
 Prescriptive analytics: find the best sets of pricing and advertising to maximize sales revenue
 Descriptive Statistics Forecasting Management science
Sampling Time series Linear programming
Mean/mode/median Casual relationships Sensitivity analysis
Standard deviation Goal programming
Range and variance Data mining Non-linear programming
Stem and leaf diagram Cluster analysis Optimization
Histogram Multiple regression Simulation modeling
Quartiles Logistic regression Heuristics
Frequency distributions Text mining Transportation/logistics

Diagnostic/Predictive
Descriptive analytics Prescriptive analytics
analytics

Databases and data warehousing

Relational database modelling

Structured Query Languages
Dimensional modelling
Extract-Transform-Load
Data Warehousing schemes
Online Analytical Processing
Non-structured Query Languages
Distributed file systems
Decision-making framework used to describe, analyze, and
understand how individuals or organizations make decisions.

Descriptive Analytics
 Descriptive analytics
 Reporting/OLAP
 Dashboards
 Data visualization
 Data visualization is the process of converting raw data into easily understood

pictures of information that enable fast and effective decisions.

 Visual analytics refers to the use of computer graphics to create a visual representation

of large collections of information.

 Purpose of visualization is to enable knowledge discovery.

 Data-driven actions are increasingly made without access to information provided by
traditional information presentation

 Information visualization is emerging as an important fusion of graphics, scientific

visualization, database, and human-computer interaction.

 Data visualization is used in software applications to provide an intuitive graphical

interface.

 It is applied to many areas to enable users to glean useful information from their data
for faster, more informed decision making.
 Data visualization allows users see several different perspectives of the data.

 Data visualization makes it possible to interpret vast amounts of data

 Data visualization offers the ability to note exceptions in the data.

 Data visualization allows the user to analyze visual patterns in the data.
A table is best when:
 You need to look up specific values

 Users need precise values

 You need to precisely compare related values

 You have multiple data sets with different units of measure

A graph is best when:

 The message is contained in the shape of the values
 You want to reveal relationships among multiple values (similarities and differences)
 Show general trends

 You have large data sets

 Bar Charts
 Line Charts
 Pie Charts
 Bubble Charts
 Stacked Charts
 Scatterplots
 Time is always placed on the x-axis since it continues to move
forward
 Y Variables have continuous values
 Familiar/conventional ordering among ordinals
 Good at
 Showing specific values
 Trends
 Comparison of relative point values
 Presents categorical variables
 Height of bar indicates value
 Double bar graph allows comparison
 Convey overall impression of relationship between two variables
 Used to present measurements of two variables
 Effective in identifying relationships between the two variables
 Emphasize the differences in proportion among a few numbers
 Pie chart summarizes a set of categorical /nominal data
 Too many segments can be harder to compare than in a bar chart
Consistency
What will happen ?

• Often incorporate uncertainty to help managers analyze risk.

• Aim to predict what will happen in the future.

• Uncertainty is imperfect knowledge of what will happen in the future.

• Risk is associated with the consequences of what actually happens.

Predictive Analytics

 Regression analysis

 Machine learning
 Statistical method that shows the relationship between two or more variables.
 Investigates the relationship between a dependent (target) and independent variable(s)
(predictor).
 A regression model is able to show whether changes observed in the dependent variable are
associated with changes in one or more of the explanatory variables.
 It does this by essentially fitting a best-fit line and seeing how the data is dispersed around
this line.
 Helps economists and financial analysts in things ranging from asset valuation to making
predictions
 assumptions about the data and the model itself must hold
 Linear Regression
 Logistic Regression
 Polynomial Regression
 Stepwise Regression
 Ridge Regression (L2 Regularization)
 Lasso Regression (L1 Regularization)
 Elastic Net Regression
 One of the most widely known modeling technique

 Establishes a relationship between dependent variable (Y) and one or more independent

variables (X) using a best fit straight line

 Represented by an equation Y=a+b*X + e, where a is intercept, b is slope of the line and e is

error

 There must be linear relationship between independent and dependent variables

 Best be tested with scatter plots

 Very sensitive to Outliers

 Use to find the probability of event= success and event= failure.

 Logistic regression when the dependent variable is binary (0/ 1, True/ False, Yes/ No) in

nature

 In its basic form, uses a logistic function to model the variable

 Parameters are chosen to maximize the likelihood of observing the sample values rather

than minimizing the sum of squared errors.

 It is widely used for classification problems

 Logistic regression doesn’t require linear relationship between dependent and independent

variables

 It requires large sample sizes because maximum likelihood estimates are less powerful at

low sample sizes than ordinary least square

 The independent variables should not be correlated with each other i.e. no multicollinearity

 If dependent variable is multi class then it is known as Multinomial Logistic regression

 Polynomial regression allows for a more flexible curve that can better capture nonlinear
relationships between the variables.
 The best fit line is a curve rather than a straight line
 While there might be a temptation to fit a higher degree polynomial to get lower error, this

can result in over-fitting

 Always plot the relationships to see the fit and focus on making sure that the curve fits the

nature of the problem

 Use of algorithms and statistical models by computer systems to effectively perform a

specific task without using explicit instructions

 Machine Learning algorithms are applied in data mining in areas such as prediction,

classification and clustering

 Different techniques

 Decision trees

 Bayesian/Markov models

 Support vector machines

 Regression
 Classification

 Predicts categorical class labels (discrete or nominal)

 Prediction

 Models continuous-valued functions, i.e. predicts unknown or missing values

 Typical applications

 Credit approval

 Target marketing

 Fraud detection
Helps decision makers identify the best solution
 Optimization

finding values of decision variables that minimize (or maximize) something such as cost
(or profit).

 Objective function

the equation that minimizes (or maximizes) the quantity of interest.

 Constraints

limitations or restrictions.

 Optimal solution

values of the decision variables at the minimum (or maximum) point.

 Prescriptive analytics is the area of business analytics dedicated to finding the best

course of action for a given situation

 Often used for the allocation of scarce resources

 Suggest actions to decision maker based on descriptive and predictive analytics

 Frame the problem

 Identify candidate actions and consequences of actions

 Assess value of consequences

 Suggest highest value actions

 Optimization is a combination of mathematical modeling and exact algorithms used to find

the optimal answer

 Decision variables

 These are the business questions to be answered

 Eg: How much raw material to purchase?

 Eg: How much product to sell to certain markets?

 Adhere to business realities/restrictions. Commonly referred to as constraints or bounds

 Eg: How much raw material is available to purchase?

  Examples

 Transportation: shipping goods from supply to demand points at minimal cost

 Equipment replacement: determining the optimum point in time to replace equipment

 Assignment problems: assigning staff to equipment

 Product mix, blending, and substitution

 Customer profitability and pricing

38
Example
 A courier service have to pick up multiple parcels from different
places in a city and deliver them to different places
 How can they accomplish this consuming less fuel?
 Types Of Decision Models

 Regression Analysis

 Machine Learning Concepts

 Optimization