Denodo DataFest 2016: Data Science: Operationalizing Analytical Models in Real-time with Data Virtualization
1 like336 views
The document presents a discussion on rapid, agile data strategies for enhancing analytics using data virtualization, emphasizing its role throughout the data science and analytics lifecycle. It details four phases of operationalizing analytical models: discovery, exploration, real-time operationalization, and predictive optimization, along with practical examples from industries like healthcare. The presentation aims to facilitate understanding and implementation of data virtualization in various analytical contexts, promoting collaboration and real-time data access.
Denodo DataFest 2016: Data Science: Operationalizing Analytical Models in Real-time with Data Virtualization
- 1. O C T O B E R 1 8 , 2 0 1 6 S A N F R A N C I S C O B A Y A R E A , C A #DenodoDataFest RAPID, AGILE DATA STRATEGIES For Accelerating Analytics, Cloud, and Big Data Initiatives.
- 2. Data Science: Operationalizing Analytical Models in Real-time with Data Virtualization Suresh Chandrasekaran Senior Vice-President, Denodo
- 3. Agenda 1.Data Science & Advanced Analytics 2.Role of Data Virtualization in 4 Phases 1. Discovery of Data Assets 2. Exploration of Analytic Models 3. Real-time Operationalization 4. Predictive Intervention & Optimization 3.Summary 4.Q&A
- 6. Data Science and Analytics Lifecycle Data Preparation Modeling Evaluation Exploratory Analytics Business Understanding Data Understanding Data Note: Adapted from CRISP-DM Predictive Optimization Real-time Operationalization
- 7. Definition of Data Virtualization (circa 2012) Evolution of Data Virtualization Data Virtualization’s Growing Role Data Virtualization data store 1 data store 2 data store 3 data store 4 data consumer 1 data consumer 2 data consumer 3 Data virtualization is the technology that offers data consumers a unified, abstracted, and encapsulated view for querying and manipulating data stored in a heterogeneous set of data stores. - Rick van der Lans Modern Approach (and Technology) to Enable Agile Data Access, Data Prep, Exploratory Analytics, Governed Data Services, Secure Sharing, and More … Enable Real-time and Agile IM / Analytics Lifecycle Data Virtualization Self-Service Data Data Integration Data Governance Single View Data Services “Enterprise Fast Data Strategy”
- 8. Role of Data Virtualization in Analytics Lifecycle Discover, Explore, Operationalize, Optimize
- 9. 1. Connect, Discover, Prepare Data Sources and Assets Anyone can explore any data source /asset, Get governed real-time access 100s of sources: Enterprise, Cloud, Big Data, Web, Files… +SDK Wizards – 3 clicks to virtualize Data Preparation & Integration with vast transform, DQ functions; Data Modeling top-down, source-up, import, export w/ validation Canonical entities, associations, Publish/Access in 14 formats by any user Business Glossary, Governed Info Self- Service, Data Services Marketplace Find more details at: datavirtualization.blog http://www.datavirtualizationblog.com/?s=sources 9 Connect, Virtualize w/Ease Data Preparation, Trfm, DQ Information Self-ServiceDenodo Model Bridge
- 11. Virtual Data Layer->Data Services->Marketplace
- 12. 2. Exploratory Analytics, Sandboxing and Collaboration Rapid Creation, Iteration and Collaboration with Virtual Analytic Sandboxes Logical Data Views & Marts Provisioned in minutes Analytical Functions supported via Virtual Layer Introspect from External Tools Collaboration – Share data, metadata and analytic model; Annotate add comments 12 Logical Views Relationships Collaboration / VersioningAnalytical Queries
- 13. Overall Business Information Solution & Sandbox
- 14. From Analytic Sandbox to Operational Reporting
- 15. 3. Operationalize Across Broad Use Cases and User Types Top 5 – Informational, Analytical, Operational, User Self-Service, Data Mgmt Flexible Publishing data access via 12+ protocols; Support Search, Browse, Query SQL access via JDBC, ODBC, ADO.NET; Data Services – SOAP, REST, Odata, JMS Hybrid execution for mixed workloads - dynamic real-time query optimization; push-down analytics, caching, batch Fine-grained access to view, row, column, cell, masking, Policy-based security Resource manager manages workload per business priority, protects source overuse 15 Find more details at: datavirtualization.blog http://www.datavirtualizationblog.com/data-virtualization- main-use-cases/
- 16. Streamlined Analytics in Healthcare Purchasing Rapid Exploratory Analytics (Discovery Zone) turned into Operationalized Analytics
- 17. Embedded Analytics in Operational Scenarios Single Customer View; Scripted Prompts Combine Multiple Analytic Insights Single View of a Customer Faster customer response from 6 to 4 mins with higher FCR Calls routed based on profile intelligence Embed Cumulative Analytic Insights Into Scripted Prompts Lifetime value of customer + Price Elasticity Retention Satisfaction / Social Feedback + Product Preference Upsell Improved customer satisfaction
- 18. 4. Integrated Predictive Analytics with LDW (EDW+BigData+Cloud/SaaS+Unstructured) High Performance Even When Processing Billions of Rows; Faster Time to Value Integrated analytics across entire set of data & information Move processing to the data paradigm … wherever it lives. Common understanding of business entity & lineage – promotes trust and reuse of “single source of truth” Find more details at: datavirtualization.blog http://www.datavirtualizationblog.com/physical-logical-data- warehouse-performance-numbers/ 18
- 19. Integrated Analytics in Healthcare Purchasing Moving from Siloed Analytics to Integrated Analytics Across Functional & Business Lines
- 20. Big Data /IoT Predictive Analytics 20 Predictive Maintenance IoT - Big Data Operational Data Predictive Maintenance Service for large customers and dealers Differentiator against low- cost competitors Increase services and parts sales – higher margins Future – optimize parts and services pricing
- 21. Summary: Denodo Supports Every Phase of Data Science and Analytics Lifecycle Business Understanding Data Understanding Data Preparation Modeling Evaluation Exploratory Analytics Real-time Operationalization Data Note: Adapted from CRISP-DM Predictive Optimization Understand Your Needs across entire lifecycle of data science and analytics Explore Denodo 6.0 capabilities as they relate to each phase Seek Solutions Advice from Denodo experts based on successful customer implementations Go Beyond … Experiment, Iterate, Share …fast and easy with Denodo
- 22. Q&A
- 23. Thank you! © Copyright Denodo Technologies. All rights reserved Unless otherwise specified, no part of this PDF file may be reproduced or utilized in any for or by any means, electronic or mechanical, including photocopying and microfilm, without prior the written authorization from Denodo Technologies. O C T O B E R 1 8 , 2 0 1 6 S A N F R A N C I S C O B A Y A R E A , C A #DenodoDataFest