Several local data scientists and women in related careers gathered at the University of Virginia on Friday to share lessons learned working and leading others in the field.
The panel was part of a daylong Women in Data Science event hosted by UVa’s Data Science Institute. Only about 15% of data scientists, who typically use machine learning and statistical methods to extract and interpret large batches of information, are women, according to a 2018 industry survey by Burtch Works.
“What does it mean to you to be a data science leader and how do you define success in that role?” moderator Samantha Toet asked a panel of women on Friday afternoon. Toet’s organization, R-Ladies, promotes gender diversity in computing and data science.
“One of your roles is integration, since data scientists can’t live in a corner, doing statistics,” said Miriam Friedel, director of data science for Metis Machine, a Charlottesville-based data analytics and machine learning company.
Since data science is a relatively new field and takes skill and methods from the fields of engineering, statistics, mathematics and computer science, it’s important for leaders to know how to explain their work, get others on board and remain confident in personal and company goals, the panelists said.
“Many of us get into data science because we really like the technical, the math-y, writing code,” Friedel said. “But one thing to ask yourself is, am I happy in the weeds with technical work, or do I want to have a broader vision and lead people?”
Good program and company leaders often do have technical backgrounds, said Tatenda Ndambakura, founder and developer of Shiri, an app to help African farmers manage and improve food production.
“People who have technical knowledge are better able to be program leaders because they have realistic expectations,” she said.
Julia Barnhardt, manager of data analytics at United Network for Organ Sharing, said that if women fear their technical skills are becoming rusty or obsolete, it’s easy to spend a few hours in an online course to catch up on a programming language.
Panelists also encouraged current and future data scientists to make sure they know their organization’s goals and teamwork before committing to data science projects.
Barnhardt’s network operated long before online data, she said, and put many processes and teams in place before a data analytics team could be effective, she said.
“For organizations like that, who have a longer history from before the internet, there has to be an organizational maturity before you can set up a data science team,” she said.
The women also encouraged others to be confident in their achievements and goals.
“The beauty of this discipline is that people bring various strengths to it,” Ndambakura said.