Thursday 18 May 2017

Data without analytics is data not yet realised (interview)

Dr Jim Goodnight, co-founder and chief executive of SAS, has been at the helm for the past 41 years. Under his guidance and some very astute management philosophies, the company has become the largest privately owned, pure-play company in the big data and analytics space.
And according to industry analysts’ reports, it’s also very much a leader in predictive analytics and machine learning solutions.
Throughout the Forum Dr Goodnight's name was used in revered tones. As Oliver Schabenberger, executive vice-president and chief technology officer said, “It may be SAS.com but it is really Jim.Goodnight.com – he has had such an amazing influence on the company”.
Goodnight admits SAS is now a little too big for him to totally oversee and he has selected some excellent VPs. How long will he continue? “As long as I am having fun and can make a difference,” he responded.
He is currently the head of an education task force with the chief executives of the Business Roundtable. There, he is focusing on helping children learn to read.
He was interviewed during the recent SAS Global Forum.

What are some of the biggest changes you have seen over the past 40 years?
After leaving the university environment all we hoped was to make it through the next year! That SAS has thrived and grown each year is a testament to the foresight, the corporate philosophy, and great people. In 2016, SAS made US$3.2 billion, had over 83,000 customers and over 14,000 employees.
SAS was doing analytics before it was cool. It can do so much more to effect change for good.
We are in the age of Analytics 4.0 and over 40 years we have had to reinvent ourselves on average every ten years. We started in 1969 writing in PL/1 for IBM 360 mainframes. Ten years later SAS needed to run on Vax/DG minicomputers. Ten years after that – 1989 - was the beginnings of democratisation of computing with the 16-bit IBM PC so we went there too. Now it is all about the cloud.
That meant a fundamental rewrite using C language and that was to become a real strength as it supported all platforms. One version of SAS and portability to different operating systems was way ahead of its time. It also marked the move of analytics to the data – we were able to use massively parallel PCs to give us the memory and processing capacity.
The next big change was the coming of the cloud and Viya is a cloud extension of SAS 9.4. It is for big tasks – high-performance analytics. It is both a superset and subset of SAS 9.4 – both platforms will continue to develop and interoperate but Viya allows you to use the power of the cloud – you can get up to 3 billion instructions per second and work on much larger data sets.
The next evolution is definitely about the rise of the machines. Analytics is not standing still – Machine learning (ML), Deep learning, Artificial Intelligence (AI), advances in insights, algorithmic intelligence, automation, and more.
Classic ML is data driven. Modern ML is where the algorithm acquires new skills by itself and that is where we are heading.
Oliver Schabenberger said we need to be careful not to label ML as AI. Is that your corporate direction?
Yes, our version of ML is about algorithms acquiring new skills. There is no computer powerful enough today to build the neural networks and approach AI status. Oliver called it algorithmic intelligence and that is more accurate.
As computer hardware increases in power and storage, what does that mean to SAS?
SAS has always had more capability than the machines it runs on so we welcome technology advances. Bigger, faster, more powerful hardware is good for SAS. It means we can solve bigger tougher problems. SAS capabilities still exceed even the biggest, fastest hardware of today.
The cloud is interesting as it potentially can link every computer together so bigger problems can be solved – That is Viya’s direction.
Tell me more about SAS visualisation techniques?
In the past analytical uses of data tended to be relatively flat – crosstabs and graphs could represent it.
Today data is multi-dimensional – you add your data to other sources and it is simply beyond human comprehension to think multi-dimensionally. Visualisation takes you so far but we need ML to reveal those hidden linkages and reveal insights.
Over the past 40 years we have seen visualisation go from standard X, Y charts to dashboards and scorecards but frankly these do not work for dimensional data – its big data in overdrive.
SAS uses a bucketing system to compress thousands of data points into perhaps 50 important ones that we can understand. We think that democratising visualisation techniques is where we can make the most difference to uncovering new insights – opening it up to everyone – not just data scientists.
You demonstrated on stage accessing SAS via Alexa. Is that the next logical step in analytics?
Anything that helps humans interact with data and analytics is a good step. By the way, SAS works with Siri, Cortana, Google Assistant and Bixby too. But these are nothing more than digital assistants and we think the “miss” rate is way too high for analytical use.
We plan to have our own analytics focused voice engine that will understand data, analytics and SAS.
Is SAS’s commitment to education a big overhead?
It has never been regarded as such. I think most CEOs of large companies have that responsibility. It takes the CEO to at least approve doing something like this.
We need to get people talking about analytics – data at rest is data not realised. Let’s just say that some 63 Universities now offer Masters Degrees in Advanced Analytics and 144 certifications. Between SAS University Edition and SAS OnDemand for Academics, there have been more than 1 million registrations and downloads of free SAS software for teaching, learning and research.
SAS training in Australia is the equal of any offered in the world and its certification program is recognised globally. Our commitment means that SAS is synonymous with analytics and people get well-paying jobs.
Where is SAS heading?
If analytics is the engine of change, data is the fuel. The opportunity is enormous and we need to bring analytics everywhere.
SAS has grown so much from my original vision, not in the technology sense but its uses. We can help analyse sports performance, reduce credit and financial fraud, address cyber security, solve marketing issues, reduce child abuse/neglect, use crowdsourcing (Gather IQ) to solve humanitarian issues, and one new area is Results as a Service to enable everyone to access SAS and pay for results.
The next decade of development is about machine learning, algorithmic intelligence, IoT and the masses of new data we will see generated by things like autonomous cars. I think there will be a move from using analytics in a reactive mode to a proactive mode – real-time.
Meanwhile, we will be looking 20 years ahead to begin to address emerging issues.
The writer attended the SAS Global Forum in Orlando, Florida, as a guest of the company.
                                         Email us: info@epoch.co.in

                       SAS Training & Placement Programs with Internship*:
                                       Epoch Research Institute India
                               Largest and Oldest SAS Training Institute
                                       Epoch Research Institute Links:
                                   Epoch | Facebook | Twitter | LinkedIn
                                   Epoch Research Institute India Pvt. Ltd.
                                 Ahmedabad |     Bengaluru         |  Chennai
                         +91 79 40327000|+91 80 49077000|+91 44 42641500

#SASELEARNING,#SASELEARNING,#SASONLINETRAINING,#SASONLINETRAININGFORBEGINNERS,#LEARNSASPROGRAMMINGONLINE,#SASCLINICALONLINETRAINING,#SASBASEONLINETRAINING#BIGDATASASTRAININGEPOCH,#SASBIGDATATRAINING#EPOCHRESEARCHINSTITUTE,#SASTRAINING,#EPOCHSASFEEDBACK,#CLINICALSASPROGRAMMING,#EPOCHCLINICALSASPROGRAMMING.
 

No comments:

Post a Comment

How to code in Python with SAS 9.4 by Epoch Research Institute

The SAS® platform is now open to be accessed from open-source clients such as Python, Lua, Java, the R language, and REST APIs to leverage...