2. Right Solution at the Right Time
Superior, Proven, Scalable and Patented
3. 3
Why CIOs Want Analytics by the Numbers
factor
to an
organization’s
competitiveness
12X
more profit
growth
Top performing enterprises use
5Xmore business analytics
More likely
to
outperform
their peers
by
8 out of 10
CEOs expect
complexity
to increase
Financial
performers
are
64%
more likely
to use
analyticsIBM Global CIO Survey 2015
#1
4. 4
The Siloed Analytic Process Stalls Adoption
85-90% of companies
have Big Data Analytics
as a top priority
But Only 6%
can attempt to
implement
a Big Data
Strategy
Source: Accenture Internet Insight Survey 2015
The Analytic Process
• Complex
• Siloed
• Messy
• High Risk
5. A Unified View grants Contextual Understanding
5
Iyka dS Unified View: The 5 Ws of Intelligence
Where
• Predictive Analytics
• Prediction and Trends
• Statistical Data
• Quantitative Data
Why
• Text Analytics
• Sentiment and Review
• Descriptive Data
• Conversations
Who What When
• Business Intelligence
• Price Time and Location
• Relational Data
• Reports and tables
6. Iyka dS Removes the Analytic Process Entirely
6
Capability
Value
Prediction
Siloed
Reporting
Comprehensive
Analytics
Comprehensive
Self-Service
Analytic Empowerment
8. 8
Knowledge Portability: Retains Knowledge
Data Integration
Data Warehouse+ Big Data
Existing Analytic Process
Universal Direct Connect
Knowledge Portability
+ =
Disparate Data
Data Preparation
• Questions are captured
shared and repurposed
• Platform learns over time
• Context is individually
tailored
• Add Ontologies, Dictionaries
and Translators
Maximum
50% Accurate
20%
Maximum
Productivity
Siloed
Knowledge Portability
Orchestration
10. 10
2X Comprehensive Analytic Empowerment
Compliant Method
Native Data is never manipulated
No data replication
Does not retain any source data
Analytics are processed externally
Read once performance
Individual sources can be locked
Reduces Compliance Risk
Universal Direct Connect
Sources only require read or ODBC Access
Rapid access to all sources
No cleaning burdens
Precision modeling
On the fly ETL
No integration
Utilizes Existing Technologies
Unified View
Translate any language and data
Upfront results
Access the Complete Picture
Cross walk and drill down capable
Add translators and dictionaries
Exports to Pivot and standard tables
Discovers Value
Knowledge Portability
Leverages dictionaries and ontologies
100% accuracy of all data
Descriptive Data capable
Questions can be shared
Platform learns over time
User defines context
Empowers Users
2X More Productivity2X Less Engineering
2X Less Infrastructure
2X Faster Results
11. 11
Insight, Knowledge and Collaboration
New Capabilities
• Define contextual understanding
• Scheme-less View Unification
• Risk Free with an instant RoI
• Upfront outcomes evidence
• Non-technical Self–Service
• Dictate proximity value
• On-the Fly modeling
• Link Discovery
• 100% Accurate
Endless Possibilities
• Analytic process enhancement
• Precision modeling mapping
• Big Data platform expansion
• Data pedigree retention
• Data dictionary crafting
• Validation and cleaning
• Federated search
• ETL optimization
12. The Iyka dataSpryng Journey
Comprehensive, Efficient and Precise Analytic Empowerment
13. 13
Iyka dataSpryng Steps
Source Selection
Users drag and drop disparate data sources on the fly
Knowledge Library
Stacks natural language queries to utilize concepts, descriptions,
ontologies, interpretations and dictionaries simultaneously
Self-Service Portal
Users are able to view, modify and export results on the fly
Platform Requirements
SaaS
Cloud capable
VPN access to platform
Windows credentials login
On-Premise
.NET web based
Windows SQL Server
Virtual option
Domain level access for admin
15. Iyka dS Federated Search Comparison
15
Search Engines
• Intuitive query search
• Does not require integration
• Reads unstructured data
Iyka dS Federated Search
• Intuitive query search
• Does not require integration
• Reads unstructured data
Plus
• No technical translation
• Stack thousands of queries
• Reads all data types
• 100% Accurate
17. User Defined Contextual Understanding
17
Same word 7 times:
4 definitions
• Definition 1
• Definition 2
• Definition 3
• Definition 4
18. Additional Benefit of 100% Accuracy
18
Proximity Value
• Descriptive Data in close
proximity
• Descriptive Data in closer
proximity:
• Value can be assigned
• Users can adjust values
22. In 3 months the Complete Picture answers decade old questions
thought impossible to answer
Proprietary and Confidential
22
“This is the most significant technology advance in our company’s 15-year history.”
-- Vice President & Chief, Outcomes Research
$90M in New
Patients
$120M in Patient
Programs Savings
$180M in Marketing
Effectiveness
Healthcare Patient Journey
#7 Most
Recognized
Brand in U.S.
23. Unified View provides 20X gain in 2 weeks versus a whole team year
long effort
Proprietary and Confidential
23
“For the first time we are able to mine all forms of text allowing a tremendous
competitive advantage to our Risk Management Services.”
-- SVP of Data Analytics and Risk Management
44 to 790 claims Processed 16K claim files in 2
weeks
$120M in found
money
Insurance Subrogation
#1 Third Party
Administrator
24. Worlds largest Big Data Analytic implementation of 28M dynamic
sources simultaneously in real-time
Proprietary and Confidential
24
“(5) apply approaches developed by the Recovery Accountability and Transparency
Board to spending across the Federal Government..”
-- 2014 Data Transparency Act
$706B gain in
transparency
Reporting moved from 6
months to real time
Initiated an “Act of
Congress”
Fraud and Compliance
Handled emergency
fund disbursements
26. 26
Centralized Repository Without dataSpryng
Analytic Process
Analytic Tools
Business
Intelligence
Visualizations
Reports & Outcomes
Databases
Internet &
Big Data
Translations &
Ontologies
Documents
& Folders
Infrastructure
Data Warehouse
or
ETL Lake
Central
Repository
• High data replication
• Slow to implement
• High Risk
• Complex
• High technical burden
• Requires touching data
Typical
Challenges
Integration
Modeling
ETL
Normalization
Joins
Data Mart
27. 27
Data Virtualization Without dataSpryng
Analytic Process
Analytic Tools
Business
Intelligence
Visualizations
Reports & Outcomes
Databases
Internet &
Big Data
Translations &
Ontologies
Documents
& Folders
InfrastructureIntegration
Modeling
ETL
Normalization
Joins
Virtual Cluster
• High processor Burden
• Slow to implement
• High Risk
• Complex
• High technical burden
• Requires touching data
Typical
Challenges
Gateways
Gateways
Gateways
28. 28
Iyka dS Compliant Method Removes Risk
Users
Analytic Tools
Business Intelligence
Reports & Outcomes
Visualizations
Databases
Internet &
Big Data
Translations &
Ontologies
Documents
& Folders
Sources
Compliant Method
29. Pivot or Relational Tables
29
ODBC or Read Access
Analytic Process
Analytic Tools
Business Intelligence
Reports & Outcomes
Visualizations
Databases
Internet &
Big Data
Translations &
Ontologies
Documents
& Folders
Infrastructure
• 12X faster implement
Does not require:
• Central or Virtual
Repositories
• Source data storage
• 90% productivity gain for
implementers and users
Does not require:
• Normalization
• Data preparation
• Cleaning
• Logical and Physical
Modeling
Iyka dS Universal Direct Connect Efficiency
30. 30
Advantages over Natural Language Processing
Ask a Question:
Define goals
Key Performance
Indicators (KPI)
Define approach
Continuous Integration:
Data Consolidation, Aggregation
or Integration
Text and Numeric Normalization
Model Data
Conceptual
Logical
Physical
Outcomes:
Max at 50%
accuracy
Not contextually
aware
Slow to implement
Can’t identify
Descriptive data
Iyka dS Native Data Processing goes direct from question to
native source. A unified view of all intelligence and
relationships can be discovered
without pre-conception
31. Normalization: Pre-defining words
31
The transformation of text into a specific class of
terms to incorporate more varieties of data:
Example: “Spring” can have multiple meanings so
the definition needs to accommodate and define
these meanings or it can’t be utilized
Challenges:
No single idea/method of normalization: stemming,
case folding, Boyce-Codd, semantical isolation,
lamentation, domain key, grep morphing etc.…..
Big Data is too unstructured to classify and define
Once the complexities are applied to the native
data most of the time the originating intent is
permanently lost
32. Data Modeling: Pre-defining the Relationships
32
A model on how words relate to each other:
Example: a model can be created to show the
relationship between a product cost and the
purchaser
Challenges:
No single idea/method of modeling: Bachman
diagrams, Barker's notation, Chen's Notation, Data
Vault Modeling, Extended Backus–Naur form,
IDEF1X, Object-relational mapping, Object-Role
Modeling, Relational Model/Tasmania…
Conceptual models need to be programmed and
scripted into logical and then physical models
Tremendous amount of data sprawl and complexity
as each individual has their own requirements
34. 34
Comprehensive Analytics Addresses All Segments
Source: Market Monitor: Total Data, Q2 2016
Worldwide Total Data revenue forecast by segment in
$Millions
35. 35
Proprietary and Confidential
New Capability, Profitability and Time to Revenue
Revenue Opportunities
Flexible Profit Models
Consulting and Support
Lead Generation
Risk Free Offerings
Workshops and PoCs
Instant RoI Pilots and Projects
RFI, RFP, LoE and SoW support
Knowledge Bank Sharing
Market Vertical and Use Specific
Help
SaaS or On-Premise Access
Marketing and Presentation
Material
White Label capable
36. Iyka dataSpryng Implementation Process
36
Project Stage
Workshop Stage
Program
Approval
Solution
Design
Solution
Selection
Strategic
Design
Assess
Free
Workshop
Workshop
Results
Pilot Scope Pilot Design Pilot
Implementation
Pilot Results ProjectProject Design Pilot Results
37. Insight Discovery Qualifying Questions
37
Identifying the Problem:
What questions do you think analytics can help solve?
If you can solve this problem what would the deliverables or objectives look like?
What barriers have stopped you from solving this problem?
Understanding Relevance:
How quickly do you need answers?
What is the ideal length of time you want your results in?
How do you measure a timely response?
38. FAQ
38
When you say “Analytics” what do you mean? Provide the evidence based results to enable a more efficient
analytic process and enabling self-service
Can I use a Data Warehouse? Yes it can treated as another source, but ideally the originating source holds
the most precise information
How do I get everyone to agree on what we are calling data if we aren’t normalizing? You don’t have to get
everyone to agree each individual will determine their own definition and value of words or sentiment
Do I need a Master Data Model? Data Modeling is done per individual needs on the fly
How do you handle security and compliance? LID does not touch or move data sources. Only users you
authorize can see the data
Sounds like your doing a lot, can this solution scale? Our solution scales via processing power leveraging its
superior efficiency. It is used by the largest analytic implementation to date simultaneously cross referencing
28M records from 150 desperate data sources changing in real time