Hadoop integration, fraud analysis, addressing the analytics skills gap, cloud analytics, SAS CEO Jim Goodnight talks to CBR
Analytics firm SAS’s annual Premier Business Leadership Series (PBLS) conference in Las Vegas this week sought to address several issues and opportunities presented by the increasing importance of data in business.
With more and more data becoming available both inside and outside companies, businesses are keen to use it to inform decision making and drive efficiencies, but they are also struggling to find the resources to cope with the deluge of information.
At SAS PBLS 2014, CEO Jim Goodnight and his staff made a flurry of announcements aimed at helping companies gain more value from their data, as well as illustrating the latest trends shaping the future of the analytics industry.
Here are the five key announcements you need to know about.
SAS Visual Analytics (VA) boasts several new features in iteration 7.1 that SAS hopes will elevate its tool to challenge the likes of Tableau and Qlik in terms of data presentation and provide more analytical capability.
One new addition is ‘goal seeking’, which improves VA’s forecasting ability. Business intelligence product manager Robby Powell told CBR this feature allows customers to set a goal and then watch as the tool adjusts the performance of all contributing factors to show them what would be required to reach that goal.
He said: "If you want to increase revenue, for instance, your sales force numbers may need to increase. The price you charge for your products, and your quality, may need to increase. It optimises the contributing values to show what’s needed to meet your goal."
The update also includes text sentiment analysis, not only collecting and analysing comments from social media, but adding context by labelling a comment either positive, negative or neutral, to help brands see how they or their products are viewed.
SAS saw its cloud analytics revenue grow 35% year-on-year this past year, compared with 20% year-on-year growth in 2013. While the company declined to share financial figures, its cloud portfolio is an increasingly significant part of the business.
It has opened a new building at its Cary, Carolina campus to house its cloud analytics team, and claims to have 400 cloud customers in 70 countries using SAS Cloud Analytics either in their private cloud, SAS’s cloud or Amazon Web Services (AWS).
One such customer is Lenovo, which has SAS installed inside its Amazon RedShift data warehouse environment, enabling it to generate reports and perform analytics on its data without bringing the data out of RedShift to take on-premise.
SAS also offers analytics-as-a-service, performing data integration and data quality assurance on customers’ data in its own cloud.
Data skills gap
CEO Jim Goodnight is concerned about the data skills gap in the US and Europe, claiming there it is as many as 1.5 million analytics jobs that need to be filled worldwide.
As a result, SAS has committed to improving STEM subject uptake (the sciences, maths, engineering and technology) by partnering with North Carolina State University to offer the US’s first masters in advanced analytics.
It is working with colleges and universities to offer SAS Analytics U, a free downloadable analytics software service that was released in May and has already been downloaded 115,000 times.
In the UK, its free Curriculum Pathways software is mapped to the national curriculum and is used in 400 schools to support STEM subject studies at GCSE and A level, for instance by providing interactive and visual tutorials on topics like photosynthesis.
SAS sees a growth opportunity in fraud analysis, and is working with the public sector to create predictive models to cut down on the crime.
Goodnight told CBR the fraud detection work takes place on its Cary campus on SAS’s own virtual machines, and targets government and state insurance programmes and unemployment payment schemes vulnerable to fraud.
He said: "It’s very much a growth area. So far the fraudsters have been winning, even against our own Inland Revenue Service.
"It takes maybe a month or more to analyse all the historical data to try and build a predictive model to predict which is most likely to be fraud. We assign a probability score to every return that says how likely this is to be a fraudulent return and that’s based on historical data looking at fraud in the past.
"These models can be very accurate and the really high probability ones are flagged for investigation. Any tax authority only has a limited number of auditors and investigators so we want to give them the ones that are most likely to be fraudulent. That helps them accomplish a lot more fraud revenue return with the same number of people."
More Hadoop integration
SAS is closely involved with big data store Hadoop already, but promised more integration will come soon. While there were no concrete announcements at the event, CBR learned the analytics firm will introduce a data management tool for Hadoop in the next year, as more and more customers dump their data in it.
Goodnight said: "Quite a few of our customers are already starting to use Hadoop as a storage facility to store data. You can store it on some very very inexpensive commodity machines and not have to pay the high price of the storage we have now. So we have a lot of customers that are either moving to it or experimenting with it."
He added that 500 "fairly large" customers are using Hadoop because Hadoop writes data in parallel and SAS reads in parallel, making it a fast way to backup, then analyse data.
SAS is also using its partnership with AWS to explore ways of working with Elastic MapReduce (EMR), Amazon’s web service using Hadoop to process data.
Business intelligence product manager Powell said: "We’re looking at Amazon EMR as a cloud Hadoop distribution where we can pull data from analytics, because currently we don’t have a product in that realm, but we hope to soon."