Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Revolution R: 100% R and more


Published on

Published in: Technology
  • Be the first to comment

Revolution R: 100% R and more

  1. 1. Revolution R:<br />100% R and More<br />Presented by:<br />David Smith<br />VP Marketing, Revolution Analytics<br />
  2. 2. August 24, 2011: Welcome!<br />Thanks for coming.<br />Slides and replay available (soon) at:<br /><br />David SmithVP Marketing, Revolution AnalyticsEditor, Revolutions blog<br />Twitter: @revodavid<br />2<br />
  3. 3. In today’s webcast:<br />About Revolution Analytics and R<br />What Revolution R adds to R<br />Resources for getting more from R<br />Q&A<br />3<br />Introducing Revolution R<br />
  4. 4. What is R?<br />Data analysis software<br />A programming language<br />Development platform designed by and for statisticians<br />An environment<br />Huge library of algorithms for data access, data manipulation, analysis and graphics<br />An open-source software project<br />Free, open, and active<br />A community<br />Thousands of contributors, 2 million users<br />Resources and help in every domain<br />4<br />Download the White Paper<br />R is Hot<br />
  5. 5. 7<br />Source:<br />5<br />R is exploding in popularity and functionality<br />Scholarly Activity<br />Google Scholar hits (’05-’09 CAGR)<br />“I’ve been astonished by the rate at which R has been adopted. Four years ago, everyone in my economics department [at the University of Chicago] was using Stata; now, as far as I can tell, R is the standard tool, and students learn it first.” <br />R<br />46%<br />SAS<br />-11%<br />SPSS<br />-27%<br />S-Plus<br />0%<br />Stata<br />10%<br />Deputy Editor for New Products at Forbes<br />Package Growth<br />Number of R packages listed on CRAN<br />“A key benefit of R is that it provides near-instant availability of new and experimental methods created by its user base — without waiting for the development/release cycle of commercial software. SAS recognizes the value of R to our customer base…” <br />Product Marketing Manager SAS Institute, Inc.<br />2010<br />2008<br />2006<br />2004<br />2002<br />
  6. 6. 3000+ R Packages from the Open Source community<br />6<br />Time Series analysis<br />Portfolio Optimization<br />Econometrics<br />Genomics<br />Clinical Trials<br />Bayesian Inference<br />Survival analysis<br />Social Networks<br />Data Visualization<br />Data APIs (Twitter)<br />.. and more<br />
  7. 7. R User Community<br />From: The R Ecosystem<br /><br />7<br />
  8. 8. Revolution R Enterprise is <br />8<br />
  9. 9. R Productivity Environment (Windows)<br />9<br />Script with type ahead and code snippets<br />Solutions window for organizing code and data<br />Sophisticated debugging with breakpoints , variable values etc.<br />Objects loaded in the R Environment<br />Packages installed and loaded<br />Object details<br /><br />
  10. 10. Interactive Debugging<br />One-click to set a breakpoint in an R script<br />Step in/out/over, inspect variables<br />Eliminate the edit -> browser -> repair cycle<br />10<br />
  11. 11. Coming soon: Revolution R GUI <br />11<br />Accessible<br />Powerful<br />Extensible<br />
  12. 12. Performance: Multi-threaded Math<br />12<br />Open<br />Source R<br />Revolution R Enterprise <br />1.<br />2.<br />
  13. 13. Three Paradigms for Big Data<br />Standard R engine is constrained by capacity and performance<br />Revolution R Enterprise offers three methods for big data with R:<br />Off-line: parallel out-of-memory analytics<br />Off-line, distributed analytics<br />On-line, in-database analytics<br />Hadoop<br />Netezza<br />13<br />
  14. 14. Revolution R Enterprise with RevoScaleRBig Data Statistics in R<br />14<br /><br />Every US airline departure and arrival, 1987-2008 <br />File: AirlineData87to08.xdf<br />Rows: 123.5 million<br />Variables: 29<br />Size on disk: 13.2Gb<br />arrDelayLm2 <- rxLinMod(ArrDelay ~ DayOfWeek:F(CRSDepTime),cube=TRUE)<br />
  15. 15. Example: Old Wives Census Analysis<br />15<br /><br />
  16. 16. RevoScaleR – Distributed Computing<br />Compute Node<br />(RevoScaleR)<br />Data<br />Partition<br /><ul><li>Portions of the data source are made available to each compute node
  17. 17. RevoScaleR on the master node assigns a task to each compute node
  18. 18. Each compute node independently processes its data, and returns its intermediate results back to the master node
  19. 19. master node aggregates all of the intermediate results from each compute node and produces the final result</li></ul>Compute Node<br />(RevoScaleR)<br />Data<br />Partition<br />Master Node<br />(RevoScaleR)<br />Compute Node<br />(RevoScaleR)<br />Data<br />Partition<br />Compute Node<br />(RevoScaleR)<br />Data<br />Partition<br />16<br />*Available for Microsoft HPC Server, November 2011<br />Video demo:<br />
  20. 20. Revolution Analytics with Netezza Appliance<br />17<br />More info:<br />
  21. 21. Revolution Analytics with Hadoop<br />HDFS<br /><ul><li>Connectors to HDFS and HBASE for interacting with data stores directly in R
  22. 22. Hadoop Streaming package for executing MapReduce jobs from R.</li></ul>R<br />Map or Reduce<br />Task Tracker<br />Task Node<br />R Client<br />Job Tracker<br />18<br />
  23. 23. Enterprise Readiness: Revolution R Enterprise Server<br />Multi-User Support<br />Production Applications<br />Integrate R analytics into Web based applications<br />Data Analysis and Visualization<br />Reporting<br />Dashboards<br />Interactive applications<br />Revolution R Enterprise Server with RevoDeployR<br />19<br />
  24. 24. 20<br />Deployment with Revolution R Enterprise<br />Desktop Applications (i.e. Excel)<br />Business Intelligence<br />(i.e. Jaspersoft)<br />Interactive Web Applications<br />End User<br />Client libraries (JavaScript, Java, .NET)<br />Application<br />Developer<br />HTTP/HTTPS – JSON/XML<br />RevoDeployR Web Services<br />R Programmer<br />Session <br />Management<br />Authentication<br />Data/Script<br />Management<br />Administration<br />R<br />
  25. 25. The Advanced Analytics Stack<br />Deployment / Consumption<br />Advanced Analytics<br />ETL<br />Data / Infrastructure<br />“Open Analytics Stack” White Paper:<br />21<br />
  26. 26. On-Call Technical Support<br />Consulting<br />Migration | Analytics | Applications | Validation<br />Training<br />R | Revolution R | Statistical Topics <br />Systems Integration<br />BI | ERP | Databases | Cloud<br />22<br />
  27. 27. Wrapping Up<br />
  28. 28. Why R?<br />24<br />Every data analysis technique at your fingertips<br />Create beautiful and unique data visualizations<br />Get better results faster<br />Draw on the talents of data scientists worldwide<br />R is hot, and growing fast<br />
  29. 29. Revolution R Enterprise<br />25<br />Production-Grade Statistical Analysis for the Workplace<br /><ul><li>High-performance R for multiprocessor systems
  30. 30. Modern Integrated Development Environment
  31. 31. Statistical Analysis of Terabyte-Class Data Sets
  32. 32. In-database R analytics with Hadoop1 and Netezza
  33. 33. Deploy R Applications via Web Services
  34. 34. Telephone and email technical support
  35. 35. Training and consulting services
  36. 36. 100% compatible with R packages
  37. 37. Easy-to-Use GUI1</li></ul>1 Coming Soon<br />
  38. 38. Further Reading<br />26<br /><br /><br />
  39. 39. Revolution R Enterprise: Free to Academia<br />Personal use<br />Research<br />Teaching<br />Package development<br />27<br />Free Academic Download<br /><br />Discounted Technical Support Subscriptions Available<br />
  40. 40. Thank You!<br />Download slides, replay (from Aug 24)<br /><br />Learn more about Revolution R<br /><br />Keep up to date with R and Revolution news<br /><br />Contact Revolution Analytics<br /><br />28<br />
  41. 41. 29<br />The leading commercial provider of software and support for the popular open source R statistics language.<br /><br />+1 (650) 330 0553Twitter: @RevolutionR<br />