Big Data: Emerging Opportunities and Challenges Venky Ravirala, Chief Analytics Officer MEDSEEK Sept 24, 2013 Outline • • • • • • • • • Big Data: Motivating with Healthcare Situation Three Converging Market Drivers So What, Now What? Big Data Survey Across Industries Recommendations for Getting Started Key Solution Providers: Platform, BI & Analytics Venky’s Analytics ecoSystem Vision for MEDSEEK A few broad Use Cases Appendix: Key Challenges Across Industries Healthcare Top-of-Mind Concerns Reforms & Regulation Costs & Risk (who pays and for what) Converging Market Drivers R&D (% of Total Spend) Digital Growth created Big Data http://practicalanalytics.files.wordpress.com/2012/01/datagrowth.png Converging R&D • Central to this conversation is the Patient/Consumer • Consumption ripples to multichannel digital media, devices and instruments • Innovation is amplified with Big Data & Analytic Insights Market Disruptions are Shifting Risk • Just like Wall Street, Healthcare is magnifying the financial risk • There are going to be Winners & Losers! • Irrespective of the payment model, hospitals will increasingly be evaluated and paid for by what occurs "Beyond the Four Walls" So What, Now What? How I get started? “Art of the Possible” Managing Healthcare Outcomes Treatment Managing Patient Outcomes Digital Influence Well Researching Preparing Treatment Evaluating Recovering Maintaining Entire Person Health Continuum 10 Gartner Magic Quadrants BI & Analytics Platforms Meta-data driven dimensional modeling (semistructured and structured data), visualization, and predictive analytics are a necessity 17 Cloudera’s Apache Hadoop 18 Who is Cloudera? The #1 commercial and non-commercial Apache Hadoop distribution Who is Cloudera? Complete, Integrated Hadoop Stack Helps organizations profit from all their data Largest contributor to Hadoop ecosystem Provides the most widely used open source distribution File System Mount Workflow APACHE OOZIE Develops the most sophisticated Hadoop operations software Trained the largest number of Hadoop Scheduling APACHE OOZIE APACHE PIG, APACHE HIVE APACHE FLUME, APACHE SQOOP Developers and Administrators 19 SDK HUE Languages / Compilers Data Integration Supports mission critical Hadoop clusters UI Framework FUSE-DFS ©2011 Cloudera, Inc All Rights Reserved Confidential Reproduction or redistribution without written permission is prohibited HUE SDK Metadata APACHE HIVE Fast Read/Write Access APACHE HBASE Coordination APACHE ZOOKEEPER Specialized Analytics Vendors 20 Analytics ecoSystem Vision BI & Analyses Portals & Kiosks Exec Dashboards Scenarios & Strategy Analytics Designs + Decision Sciences Strategy & Decisions Simulations s Tests ytic Ana l IT & Data Systems Decision Support & Strategy Data Models /Cubes Diagnostics & Quality SAS/R ETL Reports Computer Technology + Info/Data Sciences Crude Information & Reports Visualization Net/Java Sharepoint Workflow Optimization Intranet Views/Marts Optimization Models Statistical Models Biz Apps EDW Wiki Hadoop APIs Meta Data Services Web Services Mobile & Social Media Adapters BIG DATA = MORE CRUDE Unstructured Data Source System DBs 21 MEDSEEK Analytics Cloud & Portals Predict Convert Empower Navigate Intelligent Version of Influence Analytic Models Layer Enterprise Data Model Patients Behavior & Propensity Insights Layer Data Access Layer Advanced Personalization Influence Strategy Responses Regional Trends Episodic Variance & Trends Hospital Encounters Rx Regimen & Gaps Encounters Call Center Lists Marketing Lists Campaigns Eligibility Medical Claims Rx Claims Lab Claims Patients Prospects PDW with MDM Data Movement Hospital Data Experian Claims Member Eligibility Disease & Wellness Data Sources Instruments Clickstream Social Media Mobile Call Center 22 Analytics ProgressionatePlan gy tr S & t r po p Su n o isi c De Dashboards Biz Apps Portal/Kiosks Reports A & BI Marts/Cubes EDW s e s ly na Integrated Marketing Unstructur ed Big Data 23 A few broad Uses Cases • Needle in the Hay Stack (Search for Nuggets) • 360-View for holistic assessments (Wallet Share) • Mine the relationships and patterns (Amazon, Genetic Factors for Diseases) • Exceptions and Outliers (Fraud, Security) • Advanced Predictive & Optimization Analytics (Marketing Mix, Portfolio Risk, … ) 24 360-View + Search + Patterns…… Medication Non-Adherence Impact http://viewer.zmags.com/publication/644753a2#/644753a2/25 25 Relationships & Risk Modeling Hospital Readmission Risk JAMA 2011;306(15):1688-1698 doi:10.1001/jama.2011.1515 26 Strategic and Personalized Outreach AFFECT PLAN PREDICT Hospital Staff Perspective 27 In Conclusion Digital Influence • Three Converging Market Drivers will cause Massive Disruption • Healthcare will go beyond managing patients within four walls to Influencing Health Continuum • Big Data Analytics will drive Advanced Personalization 29 Appendix 30 Key Challenges Across Industries 31 Key Challenges Across Industries 32 ... customers’ offline experience and marketing using location • Internet and Mobile companies focus on behavior • Monitoring and Security are big operations • Big Data Analytics is typically a separate... Precision Marketing Self-Management Tools Compelling Calls to Action Targeted Content Advanced Personalization Any Questions Before we jump into Big Data Across Industries? Big Data Survey Across... suitable technologies, select options and prove the concept through ROI • Achieve organizational understanding and work from a roadmap • Expand the talent pool: IT, Data Scientists, Analytics, Business