Part 2.

Identify research questions

Tasks
Define Business Identify and
Problem or Understand the
Opportunity Stakeholders

Assess Current Define Future

State State

Formulate Plan Business Data

Research Analytics Approach
Questions
Define Business Problem or
Opportunity
 Business
Excessively high

Problem or overhead
Businessfor requests
problem
processing

Opportunity
Will Will new
What are the reengineering
Research Research software
Research
A higher-level description than a reasons for
question 1
the process
question 2
decrease the decrease
questionthe
3
overhead?
research question/problem overhead? overhead?
Often at the start of an analytics initiative the
task of identifying a problem is not given
enough attention or the right stakeholders are
not found. This leads to:

1/ Tendency to jump into a

solution mode without
understand of what needs to be
done and why

2/ Problems not being analysed

enough

3/ Misdiagnosis of the business

problem
Two processes happen in parallel
Business problem discovery to familiarise with Identifying and understanding the right
the subject matter and ensure the right research stakeholders to ensure the needs of the
questions are asked. interested parties are taken into account during
the research process.
Technique: Business model canvas
A business model canvas describes how an enterprise creates,
delivers, and captures value for and from its customers.

It can be modified to include a data view on how data is used

by the enterprise at a high level.

— A Guide to the Business Analysis Body of Knowledge® (BABOK® Guide),

IIBA
Key Key Value Customer Customer
Partnerships Activities Proposition Relationships Segments
Value-add and Customer acquisition
non-value-add and retention

Key Channels
Resources
Who are the third Products or Group customers
parties that support Physical, financial, services that solve Comms and delivery with common needs
the business intellectual, human customer problems oriented and attributes

Costs Revenue
structure Streams

Every entity, product, or activity within an enterprise has

an associated cost. One-time purchase or ongoing subsription
Usage considerations
Strengths: Limitations:

● It is a widely used and effective framework ● Does not account for alternative measures
that can be used to understand and of value such as social and environmental
optimize business models impacts
● It is simple to use and easy to understand ● The primary focus on value propositions
does not provide a holistic insight for
business strategy
● Does not include the strategic purpose of
the enterprise within the canvas
Identify and Understand the
stakeholders
 Each stakeholder group:

Identifying and ❖ articulates different needs and

objectives
understanding
stakeholders allows to ❖ poses different types of research
questions
actively engage and
❖ is interested in different volumes
collaborate with the right
and timings of analytics results
people.
❖ holds different skill sets
❖ possesses different levels of
experience with analytics
It is critical to understand the custodians and consumers of
data to identify the relevant stakeholders for an analytics
initiative.
Technique: Stakeholder analysis
Stakeholder analysis is performed in
several steps:
1. Identify the list of stakeholders
2. Understand their roles in the
initiative
3. Understand what the connections
are between them
4. Understand their attitudes
towards the initiative
5. Find out if they have decision
making authority
6. Understand their influence
Stakeholder list
A stakeholder list is the easiest way to start the analysis.

It is a structured list with defined attributes that you use to capture stakeholder information.

Name Title Email/Phone Department Project role Involvement with

data
 Organisational CEO office
chart
An organisational chart is a way
Production Project
to visually display the formal dept
IT dept
management
relationships between
stakeholders. Service
Quality PMs
desk

It will typically show the Line 1

Develo
BAs
pment
departments/positions and
reporting lines between them. Line 2 Testing
 Onion diagram
It lets you map the involvement
of the stakeholders in your
initiative.

The rule is: the further the

stakeholder is from the center,
the less involved they are.
 Stakeholder
matrix
Once you understand your
stakeholders’ involvement, you
can start analysing the impact
and planning the collaboration
approach.

A stakeholder matrix is a good

framework to do so.
Assess current state
Understanding the current state of the organization or context
of the proposed change is fundamental to informed decision
making
Technique: Process analysis
Process model Visitor submits a
claim

System defines a
priority

no yes

Is it critical?
Emergency
Claim manager
response manager
processes the claim
processes the claim

System sends the

result to the visitor
To create a process model, analyse the
documentation or talk to the stakeholder and
do the following:
1/ Identify the outputs

2/ Identify the steps to get to those

outputs

3/ Identify who does the job

4/ Identify how the job transitions

between participants

5/ Identify the triggers to start the job

Enrich your
model Visitor
Claim
manager
System
Emergency
response
Submit
1. Add swimlanes a claim
Define
2. Add external data priority
sources / dependencies no yes

3. Expand subprocesses Process Is it critical? Process

4. Add text description for the claim
+
the claim
+

the steps
5. Add RACI for
Send the
participants email
result
Technique: Root cause analysis
All the tools, techniques, and models used for current state
analysis should be evidence-driven.
 Root cause
analysis
Observed
symptom
Category 1

Root cause analysis is used to

Category 2
identify and evaluate the Cause Cause
underlying causes of a problem.
Category 3
Cause

1. State the symptom

2. Identify categories of causes Category 4 Cause
3. Identify causes for each category
4. Perform impact assessment
5. Decide which ones you can/want Cause

to solve
Technique: Exploratory data
analysis - part 1
Investigate, analyse, and summarise data to uncover patterns.
 Exploratory analysis is performed in

EDA Schema several steps:

1. Check consistency and integrity
2. Review descriptive statistics
3. Verify interdependence and
collinearity
4. Formulate missing data
5. Provide visualisation
6. Conduct feature engineering
7. Build and test basic hypothesis
8. Refine research problems based
on insights
9. Report the initial finding
Visuals are the primary vehicle through which
the data is understood in exploratory data
analysis. Typical visualisations include:
1/ Univariate plots

2/ Bivariate plots

3/ Special purpose plots

 Basic descriptive parameters
179, 180, 160, 168, 170, 180, 180, 155, 185
Range: max_value - min_value = 185 - 155 = 30

Mean: summ / count = (179 + 180 + 160 + 168 + 170 + 180 + 180 + 155 + 185)/ 9 = 173

Median: 155, 160, 168, 170, 179, 180, 180, 180, 185
Median = (count + 1) / 2

Mode: max_count_of_a_value

count(155) = 1
count(160) = 1
count(168) = 1
count(170) = 1
count(180) = 3
count(185) = 1
Basic descriptive parameters
 Basic descriptive parameters
179, 180, 160, 168, 170, 180, 180, 155, 185
The Standard Deviation is a measure of how spread out numbers are.
Variance which shows how much the numbers may differ from the mean.
Variance = avg(squared differences from the Mean)
Mean = 173
Sum of squared diffs = (179 - 173)2 + (180 - 173)2 + (160- 173)2 + (168 - 173)2 +
(170 - 173)2 + (180 - 173)2 + (180 - 173)2 + (155 - 173)2 + (185 - 173)2 = 36 + 49 +
169 + 25 + 9 + 49 + 49 + 324 + 144 = 854
Variance = Sum of squared diffs / Count = 854 / 9 = 94.(8)
The Standard Deviation = Square root of Variance = sqrt(94.(8)) = 9.74
Technique: Exploratory data
analysis - part 2
Correlation
 The Pearson correlation is the most commonly used measurement for a linear relationship
between two variables.

r = 0.995 r = -0.972
Linear regression is a data plot that graphs the linear relationship between an independent
and a dependent variable.

Y = 3.69*X -9.59 R2 = 0.992

Technique: Exploratory data
analysis - part 3
Technique: Exploratory data
analysis - part 4
Histogram
Bar chart
Column chart
Column chart (stacked)
Column chart (stacked, normalised to 100%)
Pie chart
Sunburst chart
Treemap chart
Line chart
Box plot (candlestick plot)
Scatter plot
Bubble chart
Heat map
Radar chart (Spider web chart)
Bullet chart

* Source: wikipedia
Case study: Anscombe’s quartet
Define future state
 Defining the future state includes
ensuring:
❖ the future state is clearly defined
Defining the future state
❖
creates a vision of the it is achievable

desired outcome of the ❖ key stakeholders have a shared

vision
change
❖ measurable objectives are
established
 Only using appropriate
measures of success can
you decide if an analytics
initiative was a success
or a failure
 Another important
aspect of defining the
future state is
establishing the scope
for the analytics effort.
The desired output from defining the future state is a clear
understanding of the business objectives and the value the
business is seeking to obtain from the analytics effort.

— Guide to Business Data Analytics, IIBA

Formulate research questions
Analysts require the skills to identify the right problem or
opportunity and to focus the team on the right question to
ensure the analytics work is guided properly.

— Guide to Business Data Analytics, IIBA

Technique: Hypothesis testing
Hypothesis testing is an act in statistics whereby an analyst
tests an assumption regarding a parameter.
Scientific hypothesis
should be falsifiable
We test a hypothesis by evaluating a random sample
of the entire population. Practically, we test two
hypotheses:

1/ The null hypothesis (H0)

2/ The alternative hypothesis (HA)

 ❖ state the two hypotheses
Hypothesis formulation
❖ formulate an analysis plan
and testing process
❖ carry out the plan
❖ analyse the results
 A p-value measures the significance of the evidence in
support of the null hypothesis. It is the probability of observing
the test statistic with the assumption that the null is true.
Plan business data analytics
approach
 When planning a business data
analytics approach, analysts:
❖ determine the capabilities and
capacity of the organization to
perform analytics
Planning the Business
❖ identify “quick wins”
Data Analytics Approach
❖ determine the type of analytics
defines how analytics questions being asked for
work will be performed.
❖ maintain traceability of business
needs, objectives, research
questions and their sources