Analytics, Business Intelligence, and Data Science - What's the Progression?

The First Step in Information Management
looker.com
Produced by:
MONTHLY SERIES
In partnership with:
Sept. 7, 2017
Analytics, Business Intelligence and Data Science:
What's the Progression?
Sponsored by:

Topics for Today’s Analytics Webinar
 Defining Business Intelligence (BI), Analytics and Data Science
 Differences in Architectures
 When to Use Analytics, BI and Data Science
 Evolution Between Analytics, BI and Data Science
 Key Take-Aways
 Q&A
pg 2© 2017 First San Francisco Partners www.firstsanfranciscopartners.com

Why is Today’s Topic Important?
 Organizations struggle with where to place and how to
manage the use of data.
 The addition of powerful analytics just adds another item to the stack of data
usage that needs to be managed.
 Organizations need to be clear about where the capabilities lie – and who is
responsible for successful application of all the varieties of using data.
 There are numerous alternatives, and there is no one reference model.
 Too many organizations are going the self-service route and are failing at
meeting their data needs.
 Without a good understanding of what will work in your organization, you
are at risk.

www.firstsanfranciscopartners.com
Defining Business Intelligence (BI), Analytics
and Data Science

Definitions
 No solid demarcation between these “styles” of using data

Definitions
 Business intelligence ‒ *an umbrella term that includes the
applications, infrastructure and tools, and best practices that
enable access to and analysis of information to improve and
optimize decisions and performance.
Then does it include analytics?
*Source: Gartner IT Glossary

Definitions
 Business Intelligence - *an umbrella term that includes the applications,
infrastructure and tools, and best practices that enable access to and
analysis of information to improve and optimize decisions and performance.
 Analytics ‒ *Analytics has emerged as a catch-all term for a variety of
different business intelligence (BI)- and application-related initiatives. For
some, it is the process of analyzing information from a particular domain,
such as website analytics. For others, it is applying the breadth of BI
capabilities to a specific content area (for example, sales, service, supply
chain and so on).

Definitions
 Business intelligence - *an umbrella term that includes the applications, infrastructure and tools, and best
practices that enable access to and analysis of information to improve and optimize decisions and
performance.
 Analytics – *Analytics has emerged as a catch-all term for a variety of different business intelligence (BI)-
and application-related initiatives. For some, it is the process of analyzing information from a particular
domain, such as website analytics. For others, it is applying the breadth of BI capabilities to a specific
content area (for example, sales, service, supply chain and so on).
 Data science – (grouped in with Advanced Analytics definition) *the autonomous or semi-autonomous
examination of data or content using sophisticated techniques and tools, typically beyond those of
traditional business intelligence (BI), to discover deeper insights, make predictions, or generate
recommendations. Advanced analytic techniques include those such as data/text mining, machine learning,
pattern matching, forecasting, visualization, semantic analysis, sentiment analysis, network and cluster
analysis, multivariate statistics, graph analysis, simulation, complex event processing, neural networks.

Definitions
 Business Intelligence
 Analytics
 Data Science
Which one(s) do I use?

Differences in Architectures

Architecture Drivers
Backward-looking Forward-looking
Operations and “better decisions” Data science, new insights, strategy
Steady state of sources Dynamic Sources
TRADITIONAL CONTEMPORARYBUSINESSGOALSAND
INFORMATIONREQUIREMENTS
TRADITIONAL DIFFERENCES
AUDIENCE
DATA SOURCE DIFFERENCES
BUSINESSINTELLIGENCE
ANALYTICS&DATASCIENCE
Quality, reliability and precision Enablement not control
MANAGEMENT & GOVERNANCE

Architecture Differences
ETL, EAI, data quality Anything that works
Suitable for batch or near-time Streaming and high volume
Queries have limitations Tuned for huge volumes
TRADITIONAL CONTEMPORARY
INFRASTRUCTURE
DATA VOLUMES
BUSINESSGOALSAND
INFORMATIONREQUIREMENTS
BUSINESSINTELLIGENCE
ANALYTICS&DATASCIENCE
DATA USAGE

It’s a Continuum
 Effective use means
exploiting data assets
 Various standard
architectures are
presented to allow for
understanding;
the reality is no single
style of architecture can
address all situations
 Note that algorithms and
query complexity are not
called out, because you can
run complex algorithms
against anything
Reports
Dimensional
Query
Predictive
Modeling
Scenario-
based
Forecasting
Goal
Seeking
Models
Normalized
Data
Structures
EDW and
Marts
Load
Hadoop
Schema on Read
Hadoop
Structure and flexibility
Sourcing and data types

FSFP Reference Architecture – Abstract
Data Insight Architecture
1
Data
Movement/
Logistics
Context
Monitoring
Controls
Management Layer
Metadata, Lineage, Work Flow, Models, Reference Data, Rules, Canonical Data
Data Access Layer
Reports, Visualization Visualization, Prediction, “Closed Loop,” Edge Analytics
Traditional Area
ERP
CRM
Finance
Traditional Data
Collection
EDW
Data Marts
Contemporary Area
Edge Processing
Ingestion
Smart Machines
Social
Bots
Business Strategy
Data
Scientists
Traditional
Stakeholders

Principles
 Determining which part of the data use spectrum to use is
a function of principles
 Most organizations just declare an architecture; e.g., “We need a lot of data
so it has to be a data lake.”
 Principles to apply:
− Architectures to deliver BI and analytics need to reflect business needs
− Supporting organizations around BI and analytics need to reflect true self service
− Final architecture solution must be based on support of both modes or vintage
and contemporary environments

Other Decision Factors
 For BI, answer these questions as YES
− Are the results intended to be repeatable?
− Will the result be made operational?
− Are you using the result to make decisions or monitor progress?
 Analytics and Data Science is more variable
− What is the level of experimentation?
− Is AI or machine learning involved?
− Are there algorithmic models involved?
 Other questions to consider
− Does any of the data leave the organization?
− What are the regulatory constraints?

An Engineering Process to Define the Correct Architecture
 Break all the previous
notions. Remember there
is no mandate for any
particular architecture, like
Data Warehouse, Data Mart,
Operational Data Store and
a Data Lake.
 Any combination is possible,
as long as it meets business
needs.
17
Understand
Business
Strategy
and Goals
Determine
Needs for
Operations
and Mgmt.
Determine
Data Needs
for Planning
and Analysis
Determine
Org.
Support
Develop
Best-Fit
Architecture
© 2017 First San Francisco Partners www.firstsanfranciscopartners.com

Organizational Considerations

What’s the Progression?
Data analytic tools and approaches go in and out of favor.
CENTRALIZED
DE-
CENTRALIZED
Data Science
• Focused group of advanced
data processing
Business Intelligence
Competency Center
• Centralized capability
to enable efficiencies
Analytics Workbench
• Facilitation of self-
service analytics via
centralized toolset
Self-Service Business
Intelligence
• Shifting greater
flexibility to the user
Business-driven Analytics
• Purchase and
implementation via
cloud, independent of IT
Citizen Data Scientist
• More automated, visual
data processing enabling
broader adoption
TOOLS APPROACHES

Organizational Drivers
IT Driven
Business Used
Business Driven
IT Supported
STRATEGY
Regulatory
External / Reputation
Internal Experimentation
Chaos Drives Innovation
RISK TOLERANCE
Specialized
Hard to find
More General
Cross-functional
SKILLS
Centralized Decentralized

Organizational Principles/Decision Factors
 Business volatility/variability
− How frequently does your business change?
 Skills
− How adaptable are your people?
 Alignment
− How well do you collaborate across functions?
 Regulatory requirements
− How tightly does your data need to be controlled?

Evolution Between BI, Analytics and Data Science

Hierarchy of Data Use Solutions
AI
Operational
Analytics/CEP
Analytic
Experimentation
What If
Investigate
Operate
Monitor
DESCRIPTIVE
DIAGNOSTIC
PREDICTIVE
PRESCRIPTIVE
HINDSIGHTINSIGHTFORESIGHT

Summary: When to Use One or the Other
USAGE
BASIS
WHAT
HAPPENED?
WHY DID IT
HAPPEN?
WHAT WILL
HAPPEN?
MAKE IT
HAPPEN BY
ITSELF
WHAT DO I
WANT TO
HAPPEN?
WHAT
SHOULD WE
DO NEXT?
PERCEIVED
MATURITY
REPORTING ANALYZING PREDICTIVE OPERATION
-ALIZE
ADAPTIVE FORESIGHT
SOLUTION
CATEGORY
CAPABILITY
REPORTING BUSINESS INTELLIGENCE INITIAL
ANALYTICS
ADVANCED ANALYTICS /
DATA SCIENCE
SURVIVAL/
OPERATE
OPERATE/MANAGE MANAGE/
PLAN
ANTICIPATE/AUTOMATE

What Comes Next

Data Science Enables the Future of Analytics
User 1.0
To each their own..
•Persona
•Tools/Language
•Containers
Data 1.0
Fragmented Datasets
•Isolated controls
•Orphaned Models
•Access patterns
Technology
1.0
To each their own…
•Analytical
Tools/Algorithms
•Visualization Models
•Platforms –
exploration to
deployment
User 2.0
Self-service power
persona
Data 2.0
Integrated, secure,
logical data
warehouse
Technology
2.0
In-place
Analytically complete
Platform virtualization
Analytics 1.0
Aggregate
Dashboards/BI
Analytics 2.0
Connected/Mashed
Datasets
Analytics 3.0
Analytics-in-place at
Scale
Analytics 4.0
Cognitive/Multimodal
Insights; Deep Learning
Hypothesis testing
Rapid Experimentation
In-situ/CEP Insights
Artificial Intelligence
The rise of Deep Learning
Source: May 2017 DIA webinar (Data Scientist interview)

Key Take-Aways
 There are many definitions for BI and Analytics.
‒ Your environment to deliver data will never fall into one single
definition.
 The architectures for delivery will vary widely over time within a
single organization.
– Focus on fit for purpose.
 Use a formal process to determine where and how the data supply
chain is sourced, executed, managed and supported.
‒ Do not adopt external reference architecture without an alignment
exercise.
 BI, Analytics and Data Science will continue to evolve.
– Don’t be afraid to “fail fast” within a comfortable cost structure.

Questions?
MONTHLY SERIES

Thank you for joining – thanks, also, to
Looker.com for sponsoring the webinar.
Please join us Thursday, Oct. 5 for the
“Data Lake Architecture” webinar.
Kelle O’Neal @kellezoneal
kelle@firstsanfranciscopartners.com

Analytics, Business Intelligence, and Data Science - What's the Progression?

Recommended

More Related Content

What's hot (20)

Viewers also liked (15)

Similar to Analytics, Business Intelligence, and Data Science - What's the Progression? (20)

More from DATAVERSITY (20)

Recently uploaded (20)

Analytics, Business Intelligence, and Data Science - What's the Progression?