The accelerated adoption of AI and machine learning, combined with the accessibility of open source software has data scientists churning out more analytical models than ever.
However, there hasn’t been a corresponding increase in business value since few models make it out of the lab and into production, observes SAS.
The analytics provider says it aims to solve this business challenge through the release of SAS Open Model Manager, which will help organisations operationalise open source models and put their data to work for smarter, faster business decisions.
Available in November, the solution offers seamless integration with Python and R, says SAS, in an announcement during its annual Analytics Experience in Milan, Italy.
SAS says the solution will bring together data scientists and IT/DevOps to help organisations register, deploy and monitor open source models in one central environment.
Users can compare and assess different models, manage champion and challenger models, and access built-in performance reports to quickly evaluate whether to retrain, retire or develop new models.
SAS says many organisations struggle to complete the last mile of analytics, partly because of cumbersome manual processes and inconsistent collaboration between IT and business users.
The burden of moving models from development to deployment is significantly eased by improving model development, production and automation.
It points out an IDC survey which noted that less than half of organisations can claim that their analytical models are sufficiently put to work, with only 14 per cent of respondents saying that the output of data scientists is fully operationalised.
We use analytical models to make the right offers to our 65 million customers... and to drive better and faster business decisions
SAS says the Open Model Manager will help organisations streamline the process for analytical models to go quickly from the lab into production, and closely monitor and revalidate the performance of these models.
“Organisations have a good handle on building and training analytical models, including open source ones, but there is often a gap when it comes to operationalising those models and pushing them into production, and a lot of the work done by data scientists is lost,” says Chandana Gopal, research director, business analytics at IDC.
“There is a need in the market for a new generation of model management solutions that allow data scientists to develop models in any language of their choice, and to properly catalogue and deploy their analytical models. With this capability, organisations can harness the value of their analytical assets and improve transparency through continuous monitoring,” says Gopal.
SAS says simplified publishing and scoring steps provide flexibility to deploy models with just a few clicks, both in batch and real time, with different operational environments.
“SAS Open Model Manager also improves governance by helping users better understand the function and performance of deployed models over time. Without the ability to continuously monitor a model’s degradation, business value and opportunity is rapidly lost,” it states.
SAS Open Model Manager will be delivered through container-enabled infrastructures, including Docker and Kubernetes. SAS says this will provide a portable, lightweight image that can be deployed in private or public clouds.
It is designed specifically to meet the needs of the open source community, no additional SAS technology is needed.
SAS says ModelOps is another key ingredient in the last mile of analytics, where organisations move models from the data science lab into IT production as quickly as possible while ensuring quality results.
The practice of ModelOps enables organisations to manage and scale models to meet demand and continuously monitor them to spot and fix early signs of degradation.
Organisations that fail to embrace ModelOps face increasing challenges in scaling analytics and fall short of the competition, says SAS.
One of its customers, Globe Telecom in the Philippines, talks about how it tackled model deployment challenges.
While the mobile and broadband provider was implementing models in both SAS and open source, its process was manual, slow and lacked governance. With SAS, Globe has dramatically cut deployment time while seamlessly working in both SAS and open source software.
“Globe uses analytical models to make the right offers to our 65 million customers, to build and strengthen our relationships with them, and to drive better and faster business decisions,” says Dan Natindim, vice president and enterprise data officer at Globe Telecom.
“With SAS, Globe analyses all the data available, including customer, billing and network data, and through SAS and open source analytical models, we work to meet each customer’s individual needs.”
Divina Paredes attended the Analytics Experience conference in Milan, Italy, as a guest of SAS.
Sign up for CIO newsletters for regular updates on CIO news, career tips, views and events. Follow CIO New Zealand on Twitter:@cio_nz
Send news tips and comments to email@example.com @divinap