As explained in detail by Michele Chambers at the IBM Netezza blog, there are two keys to getting fast performance with statistical analysis on massive data sets with R:
- Massive parallelization: break the problem down into small pieces, and run them in parallel
- Bring the R engine to the data (not the other way around), to avoid data transfer delays
In a live webinar tomorrow, Wednesday February 29 (at 11AM Pacific), IBM Netezza's William Zanine and Revolution Analytics' Derek Norton will demonstrate this principle in action, with Revolution R Enterprise for IBM Netezza. Here's the description of the webinar:
Everyone involved in high-stakes analytics wants power, speed and flexibility regardless of the size of the data set and complexity of the analysis. Trailblazing organizations that have deployed IBM Netezza Analytics with their IBM Netezza data warehouse appliances (TwinFin) with Revolution R Enterprise are getting all three. Register for this webinar to find out how.
To set the stage, we’ll provide a brief overview of the “R” statistical analysis language, the Revolution R Enterprise framework (with R at its core) as well in-database analytics on IBM Netezza Analytics Appliances. We’ll be talking about what Revolution Analytics brings to IBM Netezza, and vice versa.
Then, we’ll complete a model-building exercise from start to finish using the combined Revolution R Enterprise and IBM Netezza solution and demonstrate the performance and flexibility of the integrated offer. Join us as we:
- Begin with data visualizations and summary statistics to gain an understanding of our data
- Split the data into training and test sets for model building
- Build models and
- Measure accuracy on both our training and test set and visualize the results.
Register today to learn why leveraging the Revolution R Enterprise and IBM Netezza analytics framework delivers power, speed and flexibility in a modern analytics platform.
To register for the webinar (or if you're reading this after the live session, to find the slides and replay links), follow the link below.