What if we stopped arguing over which analytics software is best, and decided instead to use them all?
Today’s data scientists come from many different backgrounds, and they bring a wide range of skills to the job. If they have access to a variety of analytics tools, along with a system to govern and deploy models consistently, they have more options for solving complex problems.
Cleveland Clinic and Cox Automotive are two organizations that have benefited from this realization. As a result, their data science programs are thriving – and so are their larger organizations.
“We have employees who are trained in multiple languages and technologies. We want to enable people to access and use languages they’re comfortable with but using a common approach,” says Chris Donovan, Executive Director of Analytics for Cleveland Clinic.
Cleveland Clinic hopes to grow analytics maturity across the whole health care system, says Donovan. Instead of centralizing analytics skills and capabilities in one team, they’re building a broad program across the enterprise. “Having a platform that enables that is critical for analytics to be successful.”
This focus on analytics has helped Cleveland Clinic transform, along with the industry, from a focus on palliative care to preventive care. Donovan explains: "How can we move away from just taking care of you when you show up as an individual patient in the ER or the doctor’s office, to looking at a population of patients and thinking about how to prevent people from getting sick in the first place?"
Combine the benefits of SAS and open source
How can you use open source analytics alongside SAS? And use the combination to simplify and govern your analytics ecosystem? Find out in this ebook. It includes stories from customers in many industries describing how they use SAS and open source together.
Becoming 'code agnostic'
At Cox Automotive, SAS® Viya® is the glue holding the analytics organization together, says Shawn Hushman, Vice President of Decision Sciences for Cox Automotive. “It removes the different political debates between the source systems, so we can focus on the modeling itself, the versioning of the models and the delivery of the models. The open platform allows everyone to use their code, leverage open source opportunities, and it opens everyone up to new code bases.”
In particular, Hushman praises the ability to integrate disparate code, processes, and information into one hub that provides consistent delivery of information.
Hushman’s team includes people around the world who code in Python, R and SAS. “We have people programming in multiple interfaces, and they’re using different ways to collaborate on model development. They have different ways they like to publish and show the output and different ways to deploy the models. SAS allows us to stitch it all together.”
The solution allows Cox Automotive to be "code agnostic," says Hushman. Instead of debating over a preferred code base, everyone can discuss the best solution for a problem together.
“We don’t care about your code preferences,” he says. “Let’s look at frameworks that can bring real change to the organization, instead of battling over which package I’m going to use for modeling.”
With that mindset, Cox Automotive has been able to make the most of its data as it transitions offline businesses like Autotrader and Kelley Blue Book to the online world.
“SAS has the ability to bring together the data scientist community like no other solution can,” says Hushman. “We think of model management as the center of our hub, because that’s where we can be agnostic and make sure everything connects.
“Our responsibility is to deliver results efficiently and make it seamless,” he says. “We want all the goodness that comes with the diversity of different algorithms, and then we bring alignment around our delivery."
SAS has the ability to bring together the data scientist community like no other solution can. Shawn Hushman Vice President of Decision Sciences Cox Automotive
Opening analytics to executives
At Cleveland Clinic, giving more people access to analytics is also a priority. Beyond data scientists and programmers who are adept at writing code and doing advanced analytics, Donovan and other leaders at Cleveland Clinic want to make data easily available to executives and managers with drag and drop capabilities and simple interfaces.
“Our leaders may not know how to build a predictive model but they need to be able to use data to make better decisions," says Donovan. “Not everyone is a data scientist. But everyone needs to be able to interact with data at their level.”
Donovan says Cleveland Clinic is redefining what an analyst is and working to create a common entry point for all levels of users. “Before, we had data everywhere and multiple tools, but we’re trying to invert that. Instead of taking the data to the people, it’s bringing the people to the data. We believe that if we create a world-class platform, that will draw the people to the platform — which will drive consistency, build communities of practice, and link people across the organization to find standard approaches.”
Analytics is not just a capability that supports your core strategy. It has to be a core strategy of its own. Chris Donovan Executive Director of Analytics Cleveland Clinic
Opening your business strategy with analytics
Hushman also emphasizes the importance of analytics across the organization at Cox Automotive. “I prefer to view analytics as the heart and soul of our organization and the foundational element of everything we do. Analytics is improving all our products, driving new products, and growing revenue across all our product suites. Analytics doesn’t drive our business; it is our business.”
Broadening the use of analytics to support all users is more than a technology tactic. It’s a business strategy. “Analytics is not just a capability that supports your core strategy,” says Donovan. “It has to be a core strategy of its own. You need to become an analytically mature organization, and you need to be world-class in that space or people will leapfrog you.”
- Article ModelOps: How to operationalize the model life cycleModelOps is where analytical models are cycled from the data science team to the IT production team in a regular cadence of deployment and updates. In the race to realizing value from AI models, it’s a winning ingredient that only a few companies are using.
- Article The Humanity in Artificial IntelligenceCould artificial intelligence be the change agent we need to solve many problems around the globe? Read how AI could accelerate our ability to have a a positive, lasting impact.
- Article Scenario stress testing: Beyond regulatory complianceScenario stress testing offers banks a way to simulate responses to a financial crisis using a wide range of conditions and levels of severity.
- Series Meet the data scientist: Daymond LingDaymond Ling believes the right personal traits are more important than technical skills when it comes to being a successful data scientist.