Our 2019 Insight survey showed that not only are women are in the minority within data science and bioinformatics departments, they are also paid less - with the average female USA salary in data science 30% less than the average male salary. In an effort to attract an increased number of female applicants for roles within data science, we asked Tara Chiatovich to write a blog about her role as a Data Scientist and share what attracted her to a career in data.
Don’t give up the data job: What brought me to a career in data – Tara Chiatovich
I started my job as a Faculty Assistant at Harvard with my mind made up to never be a researcher. Sure, I would be offering administrative support to some amazing research minds who could mould me in the ways of inquiry. And, yes, as a Harvard employee I would qualify for 90% off tuition that I could use toward great courses on research skills ranging from applied longitudinal data analysis from the professors who literally wrote the book on it to qualitative portraiture from the woman who invented the technique.
But I knew I didn’t want to be a researcher because as an undergraduate I had done research. As someone who wanted her work to help children, research to me meant data collection with children. And data collection with children—for all their welcome cuteness and charming spontaneity—meant:
I did all of the above for what? Just to see a couple of stars signaling statistical significance in my output? Not worth it. I was not enough of a data person to be willing to do all that again. In fact, I was coming to the conclusion that I wasn’t a data person at all.
And even beyond the cost of data collection, I didn’t see any clear path between findings from data and helping children. In my own reading of other people’s research, I had come across so many amazing findings with the potential to really improve children’s academic and social skills. Like the paper from the 70s showing that pairing a socially withdrawn child with a younger peer vastly improved the older child’s social skills. That paper was great science! Yet I looked at how schools were set up—with children only ever interacting with their same-age peers—and didn’t see any evidence that the great science had been picked up by school administrators.
Even if the people who spend every day with children had my research in front of them, they wouldn’t be compelled to read it and probably would not understand whole sections of it. How could my work help children if no one who worked with children took anything from it? So instead of embarking on a career in research, I had a vague plan to use my tuition reimbursement as a Harvard employee to get a master’s degree in school counseling. That is how I would make a difference in children’s lives, not research.
For my new job, I was assigned to support Dr. Paul Harris—a developmental psychologist—and Dr. Judy Singer—a statistician. I thought that I would get a lot out of working for Paul because his expertise in child development would inform my future work with children. And maybe, maybe I’d learn something useful from Judy. But probably not because I wasn’t a data person.
Then she let me audit her linear regression class, and I was hooked. She had a way of making even complex statistical concepts not only clear and accessible but riveting. She explained residuals to us using Zagat ratings of overall quality based on price. The places where the quality was better than our model would predict based on price—the data points with a positive residual—were a good value, and those are the places she would go on vacation (because, yes, she ran regression models to inform her choices as a tourist). When we covered multiple regression—where predictions are based on not one but two or more variables—she didn’t tell us how it worked. She showed us by building a 3-D model made out of tinker toys of how both a graduate program’s size and its selectivity together made up predictions of its ranking.
And all of her individual lessons drove home two important points:
I sat through Judy Singer’s class two more times, once as an enrolled student and once as a teaching assistant. That’s how much I loved it. And then I went to Stanford to get my Ph.D. in Education, and sat through many, many more statistics classes.
Today I work for Panorama Education, an ed tech company in Boston that partners with schools throughout the country on surveys to give students, teachers, and parents a voice. Panorama also makes software that displays student survey results alongside the traditional academic metrics of attendance, coursework, and behavior. In this manner, it gives a “Panoramic” view of how students are faring.
In my role as Research Scientist, I look for trends in national student data that can inform where educators focus their attention—like the analyses I did to show which areas of social-emotional learning are the best predictors of students’ academic performance. I also inform how our data are summarized and visualized in our products to ensure that anything we show is both accurate and informative.
In all my work, I keep in mind educators who are not data people and how to best reach them—what will click for them so that the children in their schools can reap the benefits of insights from data.
Because once I thought that I was not a data person.
And then a statistics professor who knew how to present data concepts to “not data people” changed that and set me on a trajectory to work with data professionally so that I can make a difference in children’s lives.
Research Scientist at Panoroma Education
Many thanks to Tara for sharing her story about how she was attracted into a career in data. We are very interested to hear your story. What attracted you to data science? How have you developed your career? What advice would you give to women looking to start or develop their careers?
Feel free to comment on the Paramount LinkedIn page or connect with Paula Keville-Spain, Senior Consultant for Bioinformatics and Data Science based at our Boston Office in the US.