Empowering Citizen Data Scientists to Solve the AI Skills Shortage -- ADTmag

Empowering Citizen Data Scientists to Solve the AI Skills Shortage

By Richard Seeley
March 20, 2018

"As demand for [Data Science and Analytics] DSA market workers grows, this growth puts pressure on the supply of DSA talent to grow in turn. However, there is growing concern that the supply of DSA workers is lagging dangerously behind demand."~ How the Demand for Data Science Skill Is Disrupting the Job Market

A skills shortage is hampering the advance of artificial intelligence (AI), according to industry research. AI technology that empowers a rise of "citizen data scientists" may be the answer.

That's the vision of Alteryx Inc., a data science and analytics software maker based in Irvine, Calif. This month, the company released Alteryx Promote, a component of the Alteryx Analytics platform, which it says "allows both data scientists and citizen data scientists to deploy predictive models directly into business applications through an API."

Citizen data scientists are analysts with above average skills but without a formal academic background in data science, explained Ashley Kramer, vice president of Product Management at Alteryx,

"I see the citizen data scientists as this emerging group," she said in an interview. "They don't have a degree in data science but they're more advanced than your average analyst. They are people with advanced capabilities like writing scripts within Excel. They are starting to get to that next level of being able to create predictive analytics. They need a little bit of help because they're not programmers."

This is the group of what in an earlier era were called power users. Alteryx is designed to help them.

"With the data scientist shortage," Kramer said, "we provide a platform that can be used by citizen data scientists in a code-free way, which is really important as they're learning the process."

Alteryx offers a "code-friendly" way for budding data scientists to begin creating machine learning models for business requirements such as predicting and preventing equipment failures. For example, the use of predictive analysis with Internet of Things (IoT) data from sensors on heavy equipment in the field might identify a bulldozer that is likely to breakdown if maintenance doesn't bring it in for repairs.

But to do that you need someone with enough DSA skills to make use of IoT data, which is where citizen data scientists might fill the bill.

"We really help those citizen data scientists understand the steps in a code-free way at first and getting to the next step of building a model," Kramer said. "We're bridging the gap between citizen data scientists and data scientists with the coding features in our product."

The citizen data scientists do not need to know how to program in R or any other language to develop a predictive model with Alteryx, which provides a drag-and-drop interface so they don't even have to do scripting. The product walks them through finding the data they need, cleaning the data, and then pulling it into predictive models.

"When you open up our core product, Designer," Kramer said, "you drag in our input data tool and you pick the data you need. We have over 250 tools that are anything from data prep to a whole predictive suite. I see the process of the citizen data scientist connecting their data, doing the right joins, doing the data prep that they need, and then using the predictive suite to get to the place where they start building the model out."

"It's this code-free way allowing them to use drag and drop to get at their data so they can build out a model directly from our platform," Kramer added.

In a sales scenario, the machine learning application created by a citizen data scientist might be able to identify customers that are not planning to renew their subscription to a product or service based on data from a customer service database. Those customers could be singled out for special attention. For example, the predictive model can also be used to identify what may work to retain a customer such as offering them a discount if the data indicates they really like discounts, Kramer said.

Having that kind of information available to customer service representatives in a call center can provide quick decision making on a tactic that may help save a subscription.

Whether it's a construction company needing to keep all its equipment operational, or a marketing firm dedicated to customer retention, this is where AI can impact a bottom line.

"These machine learning models can be powerful in any line of business," Kramer said.