The Advanced Analytics solution space is complex and enterprise software vendors have been aggressively building and acquiring capabilities. This rapid change has resulted in the following situations: a) confusion surrounding vendor and product names, b) difficulty accessing documentation and software and c) a widespread lack of understanding of vendor-specific credential and certification programs. The objective of this article is to shed light on these three topics.

Reducing the Confusion

There are dozens of Advanced Analytics players, so for brevity's sake this post will focus on the leaders. Gartner published its first Advanced Analytics Magic Quadrant in Feb 2014, identifying SAS, IBM, Knime and RapidMiner as Leaders, StatSoft, SAP and Angoss as Challengers, and Revolution Analytics and Alteryx as Visionaries. The Forrester Wave" Big Data Predictive Analytics Solutions in Q1 2013 identified SAS, IBM and SAP as Leaders, and Tibco Software, Oracle, StatSoft and KXEN (now SAP InfiniteInsight) as Strong Performers. Let's use a simple methodology for building out a list to analyze: If either Gartner or Forrester identified a firm as a leader, we will include that firm. If a firm was included as a Challenger/Visionary by Gartner and a Strong Performer as Forrester, then we will include them as well. Finally we are going to include R, because it is a major player in Open Source and many of the other platforms build on and support R integration (including SPSS, Knime and RapidMiner). With this complete methodology we end up with the following seven products:

  • SPSS (by IBM)
  • Knime
  • R
  • RapidMiner
  • SAP Predictive Analytics
  • SAS
  • STATISTICA by StatSoft

Sharing the Resources

Our first objective is complete - we have nailed down our list of important vendors and product names. Now let's move on to our next objective: providing high quality information on documentation and software for each platform that we have chosen. Luckily all of them offer at least a downloadable trial version. Both Knime and R (our Open Source options) have an unrestricted free download. The remaining 5 products offer a timed trial download with no feature restrictions or a feature-restricted download that does not expire. RapidMiner offers both. The following table summarizes free download features by product:

If you are interested in learning more or downloading a product, see the links below. Note that the download links I have provided below were valid as of October 3, 2014 and are generally for Windows-based operating systems.

SPSS (by IBM)

User Guide

Download (Trial Version - 14 days)

Knime

User Guide

Download (Complete Version)

R

User Guide

Download (Complete Version)

RapidMiner

User Guide

Download (Two options: Limited version - no expiration or Professional version - 14 days)

SAP Predictive Analytics / InfiniteInsight

User Guide

Download (Trial Version - 30 days)

SAS

User Guide

Download (Limited version - no expiration)

STATISTICA (by StatSoft, a Dell company)

User Guide

Download (Trial Version - 30 days)

Going Legit

Alright, we are almost there. Let's complete our survey with a review of vendor-specific credential and certification programs. Again keep in mind that this information is as of October 3rd 2014. Things can shift around quickly, especially with training and credentials in a growing industry. Here goes:

IBM offers 5 SPSS-specific certifications in the Associate and Specialist programs. The IBM Certified Associate contains the SPSS Modeler Data Analysis, Data Mining and Professional certifications and the IBM Certified Specialist contains the SPSS Statistics Level 1 and Predictive Maintenance and Quality certifications.

Knime offers "Train the Trainer" certifications to trusted instructors. Knime does not currently offer any end-user certifications.

R is both a language and an open-source project. But it is not a for-profit firm and therefore does not offer a credential program. However credentials and certificates in R are available through many 3rd parties including Revolution Analytics, the University of Washington and Statistics.com.

RapidMiner has a progressive certification system with three levels (Analyst, Expert and Master). Expert requires completion of the Analyst level and Master requires completion of the Expert level. Register for remotely proctored online exams directly - completion of classes is not required.

SAP offers training on SAP Predictive Analysis, but no credential or certificate.

SAS has a mature and well developed certification system with 8 credentials spanning programming, Advanced Analytics, BI, Data Management and system administration. The most relevant credentials are SAS Certified Statistical Business Analyst and SAS Certified Predictive Modeler.

DELL has no credential program for STATISTICA, but has 3 free and 4 paid classes available. Filter "STATISTICA" because there is no dedicated training page for this product.

The following table summarizes credential depth by product:

Credential Depth by Vendor

That's a Wrap

Advanced Analytics obviously does not contain an easy-to-understand set of vendors and products. If you will be working on high stakes problems or with very large datasets you many want to make sure you are supported by a commercial entity. Although there is a great deal of overlap from platform to platform in data manipulation techniques and in statistical principles behind the scenes there can be very large differences in performance and efficiency from implementation to implementation. Advanced in-memory technologies and massively parallel operations using graphics processor cores are two examples where capabilities differ radically across products. As with all technologies the best approach is usually to experiment and learn before committing yourself to a platform. Happy modeling!