Deprecated: Function set_magic_quotes_runtime() is deprecated in /home/mwexler/public_html/tp/textpattern/lib/txplib_db.php on line 14
The Net Takeaway: CABbing with SPSS

OTHER PLACES OF INTEREST

Danny Flamberg's Blog
Danny has been marketing for a while, and his articles and work reflect great understanding of data driven marketing.

Eric Peterson the Demystifier
Eric gets metrics, analytics, interactive, and the real world. His advice is worth taking...

Geeking with Greg
Greg Linden created Amazon's recommendation system, so imagine what can write about...

Ned Batchelder's Blog
Ned just finds and writes interesting things. I don't know how he does it.

R at LoyaltyMatrix
Jim Porzak tells of his real-life use of R for marketing analysis.

 

HOW DID YOU GET HERE?

nettakeaway.com
https:
https:
https:
https:
nettakeaway.com
https:
https:
https:
nettakeaway.com

 

 

 

CABbing with SPSS · 06/12/2008 02:19 PM, Analysis

I had the honor of being invited to Chicago to provide feedback on SPSS product directions as part of their Customer Advisory Board, or CAB.

I will not be able to talk about the details of what I saw, but I wanted to talk a little bit about some our conversations to help folks understand some things about where SPSS is heading.

Big focus on enterprise integration. In the past, they really saw the products as silos, and did little to get the products integrated into the enterprise workflow. Now, they are offering products to combine the different product capabilities into a more cohesive whole. This is at the macro level (Predictive Enterprise Server) as well as on the products (unifying codebases to allow cross platform and cross-product consistencies). Clementine will continue to be the high-end offering, however, with more non-statistical stuff happening there first, as well as better handling of big data. I didn’t get to spend as much time with the Clem guys as I would have liked, so not much to talk about here.

Heard our complaints. I was pleased to have a unanimous “we agree” from other CAB participants when we complained about the lack of “proc SQL” from SAS, or the ability to run SQL-like queries against an SPSS dataset for manipulation, merges, and counts. Clearly, this is a pain the userbase has felt that SPSS just hasn’t really recognized til now. We also all complained about desktop speed, and SPSS suggested that they might have some things in mind to speed up certain parts, sort of like the multithreading they started releasing in v16. We also all mentioned the old fashioned syntax window, and they suggested that there may indeed be a pot at the end of the rainbow. They also showed us some of v17 and v18 plans, and there are some great things ahead. More modern models for the stats folks, but also an increased emphasis on making it easier for junior analysts to do basic analytic tasks. What was clever was the range of things they are helping solve, from tactical “help me do this step” (like the data deduper) to the more strategic “help walk me through how to do this multistep process to get this output”. And for some stuff we saw, we were astounded (or at least, we used very graphic language about what we saw). Yes, there are hints of what’s to come in this paragraph.

Increasing emphasis on open. While it was nowhere near what I asked for (Release an R-only front end that looks like SPSS! Allow all SPSS features to be API callable from Python!), they were certainly considering expanding the integration of Python and R with SPSS. I think this is a great approach, since SAS defines “open” as “if you do it in SAS, it’s open to other things you do in SAS”. When competing against a giant, flanking is very powerful, and I think SPSS might be on to something here. And time to start learning Python; SAX Basic will be supported for a little while longer but I think its time is drawing to a close. Besides SPSS Developer Central, the SPSS-L and SPSS newsgroups, the SPSSTools.net site by Raynald Levesque, and anything by SPSS’s Jon Peck on the web will help you get up to speed on advanced uses of Python and SPSS.

You can do a lot with the basics. One guy was reviewing all the analyses he did, from store location analysis to sales forecasting to account-rep assignments and sales force management. We started asking how he used the geo-mapping features, or how he handled seasonality as part of the neural nets, when he raised his hand to stop us. “I do all this with regression and ARIMA.” Sometimes, we love our new toys so much that we forget about the tried and true. While this guy said that he would look at these other ideas we’ve suggested, we were all thinking “perhaps we should just try some regression instead of jumping onto some pattern detection fancy thing”.

Lots still in there, stuff that neither you nor I have even looked at. Did you know that you can output to XML from SPSS script? That you can copy dictionaries from one file to another? That variables can now have “custom attributes” like tags? There are lots of new convenience features in each version that don’t really get lots of play, you sort of have to find them. If you have a ton of script to work around something annoying, might be useful to look at the latest PDFs to see if SPSS added something to solve the issue.

Shoutouts to Karl Rexer of Rexer Analytics who was great to catch up with, and Bob Muenchen, author of the best book on using R, R for SAS and SPSS users. BTW, this 80 page early version is now a 550 page book published by Springer-Verlag, pre-order at Amazon.

I hope SPSS will give me permission to talk more about some of the things we spoke about, and keep your eyes peeled to learn more about the future of SPSS.

* * *

 

Name
E-mail
http://
Message
  Textile Help
Please note that your email will be obfuscated via entities, so its ok to put a real one if you feel like it...

Admin
powered by Textpattern 4.0.4 (r1956)