I'm sure I'm not the only one who has read and contributed to threads on the internet about all the different languages used for data mining. But one aspect that's been left out of most of these comparisons is that SAS is more than a 4th generation programming language (4GL). It's always been and always will be more than a language because SAS has been engineered to be an analytic platform -- or to put it another way, an analytic processing environment designed to support the entire analytics life cycle.
What does this mean? It means that our software engineers have developed the environment to take advantage of underlying hardware such as CPUs, memory, etc., so that SAS users don't have to concern themselves with the details of leveraging the hardware efficiently. The SAS environment does it for them.
A good programmer may be able to use another language to create code that assists with efficiency, but that's complex coding that takes time, and not all programmers have that level of ability. In the long run, that can lead to inconsistencies in running and maintaining your business processes. SAS provides this type of efficiency either automatically or with a simple option setting.
It's like the choice between a potluck dinner at a friend's house versus eating out at a nice restaurant. Some of your friends will prepare better food than others, but few will provide the level of service and quality that a trained restaurant chef offers. In addition, good restaurants stand behind their food and service.
Why doesn't this topic come up more often? Probably because it's related to back-end architecture and not as easy to show off as a nice dashboard with the end results of the processing.
SAS provides either a Service Oriented Application (SOA) based architecture or a more modern cloud friendly micro-services based architecture (SAS Viya), or a combination of both, all engineered to make it easier for users to manage data, analyze it and deploy results in a consistent, governed and highly efficient manner -- regardless of the size of the data involved. If you're ready to learn more, make plans to attend Analytics Experience 2017, September 18-19 in Washington, D.C.
4 Comments
Completely agree. I've done my fair share of communicating on social sites (e.g. LinkedIn) that SAS is much more than a language and that you can't meaningfully compare it to other ecosystems such as Python or R.
In addition to sharing your sentiment of SAS being an analytic platform, I also remind people that SAS is a strong and proud company that stands behind its products. By that I mean SAS is one of the best companies at providing product support, training and customer service. You'll never find its rival in an R or Python ecosystem.
Jared,
Thank you for taking your time to read and comment on this blog. I am glad to hear you feel the same and that you too are trying to educate people on this same topic.
Regards,
David
How would you say this compared to Microsoft standing behind R server and embedding this technology into their Azure ecosystem (with associated support, training and customer service)?
SAS needs to mindful that their market dominance will not last without significant investment and innovation. They appear to be losing in the academic space to R (SAS university addition doesn't appear to be cutting it), have failed to utilise open source development communities (a licence is a barrier to entry for most), failing to adequately capitalise on the hype surrounding AI (even though they are a company that maintains my respect for continuing to present a realistic view of the state of play) and deployment and integration capabilities are lagging behind.
Rich,
I would say that SAS is mindful of the analytics market and as per our 2016 annual report: "SAS reinvests twice the percentage of annual revenue into R&D than our competitors. Last year this commitment delivered an expansion of our platform - SAS Viya ...". We will have to agree to disagree about whether the SAS University Edition isn't cutting it or not, given the number of downloads and usage statistics I think it is providing many people with a valuable "free" version of SAS that can be used to help some learn and appreciate how SAS software works to solve problems. Not sure if you are aware that SAS has its our area on GitHub and has contributed several packages such as the sas_kernel, which is a Jupyter kernel for SAS and saspy, which is a Python interface module to the SAS System which works on Linux, Windows, and the mainframe, and python-swat, the SAS Scripting Wrapper for Analytics Transfer(SWAT) package which is a Python client to the SAS Cloud Analytic Services (CAS). See https://github.com/sassoftware for more information/code. I'm glad you used the word "hype" to describe the buzz around AI, I'd expand this to include Machine Learning (ML) as well as Cognitive computing as well. SAS has been a leader in ML for years and continues to improve on and develop new algorithms within this space. For example the gradient boosting algorithm within SAS Visual Data Mining and Machine Learning (VDMML) is a great improvement other gradient boosting algorithms that are currently available on the market today. SAS Viya is a new architecture which makes it easy to integrate and deploy SAS in a variety of cloud platforms including Azure. Not sure if you are aware but SAS has been a member of Cloud Foundry, the world's leading open source platform for cloud applications. As far as deployment capabilities SAS Viya based solutions are using the same open source deployment mechanisms that others are using such as yum and ansible. And finally your comment about integration I will have to disagree with since SAS has always integrated well with other systems whether it was based on SAS 9.4 SOAP architecture or the micro-service based SAS Viya architecture. Both provides support for standard APIs integration such as JAVA and REST and others as well as data integration through ODBC, JDBC, just to name a few.
Thanks for reading and commenting on this topic.
David