Using R on your Oracle Data Warehouse

Since it is Predictive Analytics World in our backyard (or are we San Francisco's backyard...?) I figured it is well worth the time to dust of some old but important news.

With big data (should we start calling it "any data analytics" instead?) being the buzz word and analytics the key operative goal, not moving data around is becoming more and more critical to the business users. Why? Because instead of spending time on moving data around into your next analytics server you should be running analytics on those CPUs.

You could always do this with Oracle Data Mining within the Oracle Database. But a lot of folks want to leverage R as their main tool. Well, this article describes how you can do this, since 2010...

As Casimir Saternos concludes in the article; "There is a growing awareness of the need to effectively analyze astronomical amounts of data, much of which is stored in Oracle databases. Statistics and modeling techniques are used to improve a wide variety of business functions. ODM accessed using the R language increases the value of your data by uncovering additional information. RODM is a powerful tool to enable your organization to make predictions, classify data, and create visualizations that maximize effectiveness and efficiencies."

Happy Analysis!

Comments:

It would be interesting if you had an article that discussed using "R" from within the database but without using ODM. It was not clear to me what capability there is if ODM is not used. thanks, --Peter

Posted by Peter S on March 16, 2011 at 01:36 AM PDT #

Hi Peter, Currently we offer 50+ in-DB statistical functions (t-test, Pearson correlations, ANOVA) for free in every Oracle Database, so you may not need to resort to R or even data mining in many cases. Of course this all depends on the requirements. More here: http://www.oracle.com/technetwork/middleware/bi-foundation/index-092760.html One interesting aspect of leveraging data mining as the execution engine is that you get interesting performance gains. One is that you not need to move any data (which you would also get if we did R in the DB without data mining), the other one is that the data mining engine offloads to Exadata Storage Servers allowing for very fast scoring. JP

Posted by jean-pierre.dijcks on March 16, 2011 at 03:00 AM PDT #

Post a Comment:
Comments are closed for this entry.
About

The data warehouse insider is written by the Oracle product management team and sheds lights on all thing data warehousing and big data.

Search

Archives
« April 2014
SunMonTueWedThuFriSat
  
2
4
5
6
7
8
9
10
11
12
13
14
16
18
19
20
21
23
24
25
26
27
28
29
30
   
       
Today