The efficacy of any piece of modern technology is determined in large part by the state of the data that informs its creation. Engineers and developers rely on data insights extracted by analysts to guide them in building solutions that actually serve a need, and with the steady rise of artificial intelligence tools, data’s impact will soon be scale well beyond the human capacity for work.

But without a diverse set of insights to inform the products and services built for people around the world, that scale could be futile at best — and counterproductive at worst.

Data science is integral to innovation across the industries, evidenced by its rapidly growing market opportunities. But as the field expands, so do its demographic inequities. Data’s value hinges on diversity in both the sets of information that make it up and the perspectives necessary to comprehensively interpret it. Without a diverse range of insights at the table to paint the full picture, the potential for any data set is grossly limited.

Srujana Kaddevarmuth (pictured), data science and analytics executive at Accenture LLP and ambassador for the Women in Machine Learning & Data Science team in Bengaluru, has observed the unproductive effects of homogeneous data teams firsthand through leading engineering and analytics departments at some of the most influential companies in the industry. Now, the data scientist is working to diversify data collaboration.

“Data science is a highly interdisciplinary domain,” Kaddevarmuth said. “It requires people from different disciplines to come together, look at the problem from different perspectives, [and] come up with the most amicable and optimal solution.”

Kaddevarmuth spoke with Lisa Martin, host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, during the recent Stanford Women in Data Science event at Stanford University. This week, theCUBE spotlights Srujana Kaddevarmuth in its Women in Tech feature.

Discovering the diversity gap

Kaddevarmuth began her career in technology as an engineer and has since amassed more than a decade of experience in data through senior analytics leadership positions at tech giants like Accenture, Hewlett-Packard, and now Google. With a comprehensive perspective on the field over its past few years of explosive growth, the data scientist is as interested in the technical aspect of data science as she is in the work of improving its inclusion.

Kaddevarmuth’s passion for inclusive teamwork drew her to WiDS where she found a similar commitment to diverse collaboration. The initiative’s global presence impressed her, inspiring her to become more directly involved in data science equity through the program.

“I care about data science, but also about accurate representation of women and gender minorities in the space,” she said. “WiDS is creating a significant impact globally.”

In addition to improving representation, WiDS is focused on supporting new data scientists as the organization builds a repertoire of skills to bring into the rapidly growing industry. The program hosts a Datathon to encourage team building and illustrate the necessity of different perspectives for a universal solution.

“[The] Datathon … helps young data science enthusiasts hone the required data science skill sets and also helps the data science practitioners enhance and sustain their skill sets,” Kaddevarmuth stated.

Diversity in data science is critical, and Kaddevarmuth is working with WiDS to foster a community that makes inclusion an active priority. The data scientist and her team developed a WiDS workshop to address the issue after discovering that men were consistently ranking higher than women using the competitive data analytics platform Kaggle, even with a preponderance of local female talent.

“Our research indicated that men dominated the Kaggle leaderboard … for India in general, despite that region having amazing female leading scientists … with multiple patents, publications and innovations to their credit,” she said.

Women made up only 16 percent of total platform respondents in 2017, according to Kaggle’s 2017 “State of Data Science and Machine Learning” survey. Kaddevarmuth discovered that the challenge for female data scientists stemmed primarily from a lack of networking opportunities to enable collaborative problem-solving. Women were unconsciously excluded in the male-dominated environment and therefore less supported than their male counterparts in delivering solutions.

Kaddevarmuth and her team worked to combat this cycle by designing a WiDS workshop with a built-in networking focus to encourage collaboration. The program also included a mentorship element to help participants through any potential roadblocks.

“Mentors worked with the respective teams and provided them with the required guidance, coaching and mentorship [to] help them with their Datathon journey,”  Kaddevarmuth said.

‘Achieving diversity in this field is a must’

The WiDS Bengaluru event drew more than 110 people and was successful in cultivating connections Kaddevarmuth predicts will continue beyond the Datathon. The data scientist counts overall team learnings and process management among the highest achievements of the initiative.

“The entire experience of being able to collaborate, look at the problem from different perspectives, and submit the code despite a lot of challenges and navigating the platform in itself was a decent achievement from my perspective,” Kaddevarmuth said.

Improving the social elements of data analysis is as much an ethical pursuit as it as a practical business solution. Kaddevarmuth says empathy is just as important as analytics in seeing the overall data picture and solving for the challenges therein.

“Collaboration is pivotal,” she said. “You need different people to come together, look at the problem and then … solve the challenges. Different perspectives are the key to being successful in data science domain search.”

It’s no secret the industry at large is long overdue for improvements in demographic representation. Companies such as Google LLC, Amazon.com Inc. and Facebook Inc. have expressed a commitment to diversity but see relatively little return on hiring and retention initiatives designed for greater inclusion. In addition to being an ethical necessity, diverse teams have a proven net positive effect on product efficacy and the bottom line.

As the industry aims to fill its 490,000 new data science roles, initiatives like the ones spearheaded by Kaddevarmuth and WiDS are ensuring the field grows in the right direction.

“Data science as a domain is evolving at a lightning speed. We … hold the solution to almost all the challenges faced by humanity in the near future,” Kaddevarmuth said. But to come up with the most amicable and sustainable solution … achieving diversity in this field is a must.”

Here’s the complete video interview below, part of SiliconANGLE’s and theCUBE’s coverage of the Stanford Women in Data Science event:

Photo: SiliconANGLE