I’ve always had questions about how things work. I started my career as a reporter for the Wall Street Journal, making a living out of asking questions. More recently, I worked on marketing at IFTTT (If This Then That), a technology startup in San Francisco, where I analyzed user acquisition and retention – asking why users wanted our products and how to get them to engage more deeply with them.
The questions kept driving me to our data. We had a vast store of it, and I knew the answers about our users were in there. But I didn’t know how to work with it. I felt like a traveler in a foreign land, who has many questions but doesn’t speak the language well enough to ask them.
Before Springboard, I’d mostly tried learning things on my own. I must have signed up for Coursera’s “Machine Learning” course nearly a dozen times and never made it further than the fourth week. So as I considered ways to learn how to work with data, I knew that I’d need more than purely self-motivated courses.
I signed up for Springboard, half-expecting that I’ll probably ask for a refund within a couple weeks. I wasn’t completely convinced that it was worth it. Then I had the first meeting with my mentor Soups. I was a little nervous before – it was the first time I’d spoken to a real Data Scientist about what I wanted, and I started wondering what I was doing in a land of PhD’s. But my mentor Soups was nothing like what I expected. He was friendly, excited, and very, very encouraging of what I was trying to do. He made me feel like I could really do this thing.
From the beginning, Soups suggested that I put what I was learning to use on the dataset that I wanted to explore at work. It was a great idea, and I started seeing the value of Springboard almost immediately. Within the first few weeks of starting to learn R, I was running analyses that dazzled my team at work. They were impressed that I had been able to learn so quickly.
The most valuable part of Springboard was the regular weekly call. Having something on my calendar every week made sure I didn’t slack. I knew what I got out of it was entirely up to me, so I made sure I had questions to ask and work to show my mentor when we talked.
Working with a mentor also made sure that I never got stuck. When I was trying to build a scoring algorithm at work to measure different types of user engagement, I couldn’t figure out how to handle the fact that some things had much more data than others. The old doubts started creeping in – “I don’t have a PhD, I don’t know what I’m doing, should I even be trying this stuff…” I was really disheartened. But in our next call, Soups told me a couple things to look up that solved my problems! This wasn’t the only time. He did it over and over again, with just a couple right words to get me over many humps that I wouldn’t have managed on my own.
After Springboard, I built FogOrNot.com, a machine learning powered map that predicts the San Francisco fog. I built and tested my algorithm in R, and I used linear regression on data from 130+ personal weather stations to create my predictions.
That project helped me get interest from companies with open data positions. My mentor introduced me to reddit. I love the community.