Big Data Analytics

Big Data Analytics Lifecycle

Week 2 Class 1
This class
• Data Analytics Lifecycle Overview
• Discovery
• Data Preparation
• Model Planning
• Model Building
• Communicate Results
• Operationalize
DATA ANALYTICS LIFECYCLE OVERVIEW
Big Data Lifecycle Overview (1)

EMC Education Services, 2015

Big Data Lifecycle Overview - Discussion
• Which Big Data Lifecycle phases do you see the Business Analyst role
playing a part in? Why?
DISCOVERY
Discovery (1)

EMC Education Services, 2015

Discovery (2)
• Learning the Business domain
• Identifying Resources
• Framing the Problem
• Identifying Key Stakeholders
• Interviewing the Analytics Sponsor(s)
• Developing Initial Hypotheses
• Identifying Potential Data Sources

EMC Education Services, 2015

Discovery - Discussion
• Why do you think the Discovery phase is the biggest time consumer for a
Business Analyst? Give examples.
Data Mining and Modeling Implementation
Process
1. Understanding the Business Process
Analyze the series of activities or tasks that are conducted to improve
the efficiency and effectiveness of business operations to achieve
business objectives or goals.

Diagram Source: LinkedIn Learning, Business Analysis Foundations: Business Process Modeling by Haydn Thomas
Data Mining and Modeling Implementation
Process
1. Understanding the Business Process
Create a business process map to visualization the business process to
better understand the roles of various stakeholders.

Diagram Source: LinkedIn Learning, Business Analysis Foundations: Business Process Modeling by Haydn Thomas
Analyzing a Business Process
Business Process Analysis Techniques

• Gap Analysis (What information are you missing from the process)
• Value-added Analysis (Does activities in your business process adds
value to the overall business process or your organization?)
• Root Cause Analysis (Find the root cause for a problem and mitigate)
• Observation (Observe the process in real time to determine if the
process works as intended or not)
• Experience Examination (Seek the inputs of Subject Matter Experts
(SMEs) within your organization)
Data Mining and Modeling Implementation
Process
1. Understanding the Business Process

Identify the business process to be analyzed

• Processes that have direct effect on business expenses, revenue,
customers and end products.
• Underperforming processes.
Data Mining and Modeling Implementation
Process
2. Collect business process data and information
• Go through all the sources of data and information about the process
• Gather data and information about the process to unearth the issues,
and to understand the objectives, and the areas of improvement.
Data preparation, data transformation

2. Explore the dataset through visualization: Create visual

representations of the data, such as graphs, charts, and maps, in order to
understand and communicate the insights and patterns in the data.
Data Mining and Modeling Implementation
Process
3. Data Modeling

Data modeling defines a visual representation of data elements, their

structures and the relationships between them. This can be analyzed by
running it through a series of scenarios in order to understand and predict its
behavior.
Data Mining and Modeling Implementation
Process
Guarding Against COVID-19 Vaccine Hesitance: The Role of Personal Health Engagement
and Vaccine Related Attitude
Data Mining and Modeling Implementation
Process
4. Analyze the Process
Improve the business process by analyzing the information and data
collected. Analysis techniques includes; measuring the effectiveness of the
defined processes in the model.
Data Mining and Modeling Implementation Process
4. Analyze the Process

Data Modeling and Analysis Techniques

• Correlation
• Clustering
• Text analytics
• Trend Analysis
• Regression
• Outlier Analysis or Outlier mining
• Prescriptive and optimization
Data Mining and Modeling Implementation Process

5. Evaluate and refine the model

Data results and discoveries generated are analyzed against the business
objectives
Data Mining and Modeling Implementation Process

6. Deployment

The data findings should be made available for business decision-

making.
The data mining discoveries should be in a language that is easily
understood by non-technical stakeholders or laymen.
Create a plan on how to distribute maintain and monitor the data mining
report or discoveries.
Data Mining and Modeling Implementation
Process
7. Decision Making
Use your findings from the analysis to make recommendations for
potential improvements.
8. Monitor for Effectiveness
Executed Business Process should quickly move into process
monitoring.
Process in developing a Business Process
Document (BPD)

A document containing sets of guidelines and activities for a business

process.
Importance of Business Analytics Lifecycle

• Provides a clear documentation of the business process and a better

understanding.
• Make business decisions based on data information
• Identify the problems and challenges that cause delay to some business
processes
• Improve operational efficiency through training of employees in
performing task.
ACTIVITIES
DATA PREPARATION
Data Preparation (1)

EMC Education Services, 2015

Data Preparation (2)
• Source - Find
• Extract, Transform, Load (ETL) OR Extract, Load, Transform (ELT)
• Load data into a repository for further consumption (Landing Zone)
• Data Warehouse (BI) and/or Analytics Sandbox (Predictive Analytics)
• Learn about the data
• Data Conditioning / Cleansing
• Visualize data before modeling

EMC Education Services, 2015

Data Preparation - Potential Tools Used
• SQL
• Informatica
• IBM DataStage
• Microsoft Azure
• Tibco
• Etc.

EMC Education Services, 2015

Data Preparation - Discussion
• Why do you believe Data Preparation is usually the most time consuming as
many suggest this is 50-90% of the effort?
MODEL PLANNING
Model Planning (1)

EMC Education Services, 2015

Model Planning (2)
• Assess the base data model construct as well as type (Ie unstructured)
• Assess the modeling techniques skillset in the team as well as will it meet the
requirements
• Assess if one or more modeling methods will be required based on the data
set and expected outcomes (Ie Dimensional, Text, etc.)
• Research what other companies have done
• Perform an initial pass of data acceptance

EMC Education Services, 2015

Model Planning - Potential Tools Used
• R
• SQL
• Excel
• Reporting tools (Ie Microsoft Power BI, Tableau, etc.)
• Etc.

EMC Education Services, 2015

Model Planning - Discussion
• What sources would a Business Analyst research and then assist in
recommending which model to use?
MODEL BUILDING
Model Building (1)

EMC Education Services, 2015

Model Building (2)
• Built using 'training' data
• Scored against test data
• Does the model address the following:
• valid and accurate test data
• output/behavior makes sense to the domain experts
• model is sufficiently accurate to meet the goal
• avoids intolerable mistakes
• are there any false positives or false negatives

EMC Education Services, 2015

Model Building - Potential Tools Used
• R
• Python
• Matlab
• SAS Enterprise Miner
• Etc.

EMC Education Services, 2015

Model Building - Discussion
• With model building being a more technical role for Data Scientists, how do
you see the Business Analyst role playing a part in?
COMMUNICATE RESULTS
Communicate Results (1)

EMC Education Services, 2015

Communicate Results (2)
• Important to have varied messages/processes to articulate to various
audiences
• Include caveats and assumptions
• Be open about being successful or not, or to what level of success
• Important to understand if the project sponsor is aligned with the output
compared to expectations/requirements

EMC Education Services, 2015

Communicate Results - Discussion
• Why is it so important to know exactly what the project sponsor's thoughts
are on the project?
OPERATIONALIZE
Operationalize (1)

EMC Education Services, 2015

Operationalize (2)
• Communicate success more broadly across the organization
• Migrate to production
• Address any performance challenges
• Train operational resources
• Document operational processes

EMC Education Services, 2015

Operationalize - by major Role
• Project Sponsor
o Answers questions related to the business impact of the project, the risks and return on investment (ROI), and
the way the project can be evangelized within the organization (and beyond).

• Project Manager
o Determines if the project was completed on time and within budget and how well the goals were met.

• Business Analyst
o Need to ensure all is documented for operational purposes as well as potential future use

• Business Intelligence Analyst

o Needs to know if the reports and dashboard will be impacted and need to change.

• Data Engineer and Database Administrator (DBA)

o Needs to share their code from the analytics project and create a technical document on how to implement it.

• Data Scientist
o Needs to share the code and explain the model to peers, managers, and other stakeholders.

EMC Education Services, 2015

Operationalize - Discussion
• What excuses do people use to not document projects at the end? How
important do you believe these documents are? And why?
Next steps
• Read Chapter #13 'Front Room Business Intelligence Applications' in eText
'The Kimball Group Reader: Relentlessly Practical Tools for Data
Warehousing and Business Intelligence Remastered Collection'
References
• EMC Education Services, 2015, "Data Science and Big Data Analytics:
Discovering, Analyzing, Visualizing and Presenting Data"
o https://learning.oreilly.com/library/view/data-science-and/9781118876138/?sso_li
nk=yes&sso_link_from=conestoga