Master’s student Erin Grand interned this past summer with Crisis Text Line (CTL), the only 24/7, crisis-intervention text-message hotline in the country. DoSomething.org, a nonprofit that helps teenagers advocate for social change, launched CTL in 2013 amid a flood of texts from teens struggling with depression, eating disorders and other problems. So far, CTL had responded to more than 9.2 million texts.
During her internship, Grand focused on ways to recruit and retain more volunteer crisis counselors. “I used hard numbers and ethnographic research such as interviews and observations to determine [counselor] needs and then suggested solutions to solve them,” she recently wrote on her blog. By the end of the summer, Grand figured out where CTL was losing volunteer applicants and launched a new system to remove those obstacles. In August, CTL offered Grand a job as a data scientist. She is working there part-time until she finishes her master’s in data science this spring. She spoke with us recently about her internship and career transition.
What made you decide to pursue a master’s in data science?
I was in a PhD program for astronomy but after my second-year research project investigating star formation with radio telescopes I decided that I wasn’t as interested in the politics of research as I was in data analysis. Given the job market, and the years of school ahead, I decided to transition to data science.
What’s the typical day like at CTL?
Every week seems to change. Lately, I’ve been in charge of tracking counselor retention. Soon I’ll be looking at how we can recruit the best, most committed counselors. I’m using my interviews with counselors and demographic information from their applications to create a profile of the ideal candidate against which we can match applicants.
What’s your favorite type of data?
Recently, it’s the information that comes out of interviews with users. I like that data analysis involves spending time with numbers and colored-coded notes from interviews and observations.
What skills did you learn at Columbia that came in handy at CTL?
I didn’t have a great handle on Naive Bays algorithms, or working with SQL and R before Columbia. I use these tools constantly now. I also improved my data visualization skills. For a class project last spring my team and I mapped the relationship between SAT scores and household income in NYC.
What’s the best part of working as a data scientist?
Making a difference every day. The company is still new, which means its data and business infrastructure is still evolving. I love that I get to help develop these systems from the ground up. I also enjoy working with big data sets.
What skills carried over from astronomy?
I wasn’t proficient in R or Python but I knew other scripting languages. I was also familiar with writing research papers, which helped in putting together this study, Music Recommendation System Using the Million Song Dataset, for my modeling social data class last spring.
How does CTL use data to bring about change?
We use feedback from users to improve our counseling. Most conversations end with a user survey: Did you find this conversation helpful? We can use natural language processing tools to analyze those responses to see what makes a counselor effective, and incorporate that feedback into our training
— Kim Martineau and Stella Lianou