Keeping up with the

quants
Group 25 Roll No
Ashish Kumar 1601024
Mayur Patil 1601063
Rohan Choudhury 1601085
Sachin Nirgude 1601089
Yogesh Kumbhar 1601125
The Rise of Analytics and Big Data
Analytics means the extensive use of data, statistical and quantitative analysis, explanatory and predictive models, and
fact based management to drive decisions and add value.

• Gathering, Organizing, Tabulating, and Depicting data

Descriptive • Very useful but doesn’t tell why the results happened
Analytics
can be • Use data from past to predict the future
Predictive • Presence of causal relationship not always necessary to predict
classified
as • Like a prescription from doctor, they suggest a course of action
Prescriptive • Includes methods such Experimental Design and Optimization

• The term Big Data is used to mean volumes of data that are unusually large
Big Data • It’s the type of data which is mostly unstructured & need analytics tools to make sense of it

Big Data and analytics based on it promise to change virtually every industry and business function over next decade.
The potential of Big Data is enabled by ubiquitous computing and data gathering devices like sensors, microprocessors.
People who do detailed analysis of data are called quants but the book is not meant for them.
What type of Business Decisions can be made analytically?
Three analytical thinking stages

Stage 1 Stage 2 Stage 3

The Problem Recognition Step

A quantitative starts with sense, observation of events)

recognizing a problem or • Experience in the job
decision and starting to solve it. • Need for a decision or a
In decision analysis, this step is action
called framing, & its one of the • Current issues requiring
most critical parts of a good attention
decision process. There are • Building on existing theories
various sources that lead to this
first step:
• Pure Curiosity ( common
The Problem Recognition Step- Contd.

• Identifying the Stakeholders for the

analysis
• Documenting stakeholders needs
• Assessing and analysing stakeholders
interest
• Managing stakeholders expectations
• Taking actions
• Reviewing status and repeating

• Focussing on decisions
• What type of analytical story you are telling
• Scope of the problem
• Getting specific about what you want to find out
Review of Previous Findings Step
Once the problem is recognized, all the previous findings connected to it should be
investigated. This is still a step within the 1st stage because investigating previous
analysis can help analysts and decision makers think about how the problem has been
structured so far, and how it can be conceptualized in different ways.

Some methods of reviewing previous steps:

• Do an internet search for key terms related to your analysis
• Consult a statistics textbook for analyses similar to you are proposing
• Talk to analysts around your company to see if they’ve done something similar
• Check your company’s knowledge management system if it has one
• Talk about the problem with analysts at other companies
• Attend a conference on analytics to see if anyone else working on the same topic

Next course of action:

Reframing the problem – Although the analytical-solving process have been laid out as
a linear one of 6 steps in three stages, it is nothing if not iterative. Although one cannot
spend forever in re-examining each step, it’s worth some time thinking about what the
review of previous findings suggests about reframing the problem
Review of Previous Findings Step
Some methods of reviewing previous steps:
• Do an internet search for key terms related to your analysis
• Consult a statistics textbook for analyses similar to you are proposing
• Talk to analysts around your company to see if they’ve done
something similar
• Check your company’s knowledge management system if it has one
• Talk about the problem with analysts at other companies
• Attend a conference on analytics to see if anyone else working on the
same topic

Next course of action:

Reframing the problem – Although the analytical-solving process have
been laid out as a linear one of 6 steps in three stages, it is nothing if not
iterative. Although one cannot spend forever in re-examining each step,
it’s worth some time thinking about what the review of previous findings
suggests about reframing the problem
 CHAPTER 3
SOLVING THE PROBLEM
Stage in which quantitative analysis is performed and problem is solved

• MODEL- A purposefully simplified • Measuring a variable is assigning it a number,

representation of the problem • Data is just a collection of these numbers • Finding consistent patterns (relationships)
• Model building involves using logic, • Problem is 1st organized through modeling among variables embedded in the data
experience & previous findings to process into critical variables, which then • Techniques for finding patterns in data:
hypothesize your dependent & independent become data after measurement graph, percentage, mean, more elaborate
variables • 3 ways to assign measures to variables: statistical techniques
• Which variable to keep and which to • - Binary Variables • Characteristic and complexity determine
eliminate depends upon purpose of model technique to be used
• - Categorical Variables
and on whether the variable is directly
relevant to the solution • - Ordinal Variables
• - Numerical Variables

Modeling and
Data Collection
selection of Data Analysis
(Measurement)
variables
 Types of Models
3 Questions need to be answered
Q1: How many variables need to be analyzed at the same time? (Univariate, Bivariate, Multivariate)
Q2: Do we want description or inference questions answered?
Q3: What level of measurement is available in the variable(s) of interest?

More than two

Two Numerical Data Two or a few
numerical data
Variables Categorical Variables
Variables
• To relate a couple of things that • For data in nominal categories (ex- • Use of regression analysis
can be measured Numerically male & female gender) • Regression is an approach to fitting
• Use of Correlation Analysis • Use of a series of techniques for an equation to a distribution of
• Assess whether 2 Variables vary categorical data analysis data gathered in the past
together • Outputs are often displayed in • But once the model is made, you
tabular form can use it to predict the future
 CHAPTER 4
COMMUNICATING AND ACTING OF RESULTS
 The essence of this stage is describing the problem and the story behind it, the model, the data employed, and
the relationships among the variables in the analysis
 When those relationships are identified, their meaning should be interpreted, stated and presented relevant to
the problem
 The clearer the results presentation, the more likely that the quantitative analysis will lead to decisions and
actions
 The presentation of results need to cover the outline of the research process, summary of results, and
recommendations for solving the problem
 Best way to present - convene a meeting with relevant people and present them with Q&A session
- write a formal report

Steps to be followed
Understanding of How to measure Initial solution
Available Data Solution Business Impact
Business Problem business impact hypothesis
 Modern methods of communicating
results
Use of Visual Analytics (Data Visualization)

STATIC VISUAL ANALYTICS DYNAMIC VISUAL ANALYTICS

See Compare set Show rise

Examples:
See parts of Analyzing
relationship and fall of 1 • Games for communicating analytical results
of values a whole & text
among data variable in how are they frequencies and models (Beer Game to teach SCM)
points respect to related to
Bar Chart other each other • 3-D model of the graphical results that one
Scatterplot Tag Cloud could walk through and touch
Line Graph Pie Chart
• Music Videos - Self contained videos that get
Histogram
across the broad concepts behind their results
Matrix plot Phrase net
using images, audio and video
Bubble Stack Graph Tree map
Chart
Heat map

Network
Diagram
Thank You