Data scientist Anne Marie Ou works at Slalom, a Seattle consulting firm. Ou’s occupation is among the fast-growing in the United States, with a 19% growth rate from 2016 to 2026 projected by the U.S. Bureau of Labor Statistics. (The average growth rate for all occupations is 7%.)
Here, Ou answers some questions about her work.
What do you do? I take large data sets from clients to help them identify patterns and predict behaviors that are significant for their business. This is a very interactive job where we are constantly evaluating their business and communicating ideas. After I extract results from algorithms, I communicate them in terms that are applicable for each stakeholder. This can range from how to read the results of a model to defining what a business’s customer base looks like at the moment. There are a lot of extensive brainstorms, meetings and coding involved.
How did you get started in that field? My background is in chemical engineering. I used to work in production at a refinery and spent some time in supply chain planning. Although it was a really challenging career choice with growth potential, I realized I wanted to pivot my math skills and use them differently in other industries. I learned Python and completed the Data Science certificate at Galvanize.
I reached out to several researchers who would have large data to work with and found a great project with University of Washington’s Climate Impacts Group. This project was awesome since it uses large amounts of future runoff estimates to project culvert sizes needed to support salmon migration well into the future.
The analysis led to the creation of a free interactive tool to help engineers design culverts in a way that accounts for the effects of climate change. … This tool has been used by several state agencies and tribes for planning and policy considerations. Due to the success of this project … Slalom extended me an interview — and that’s how I got started with my data science career.
What’s a typical day like? There is no typical day in data science consulting. I usually travel three to five days a week to be on-site for client meetings and brainstorming sessions. Frequently, I spend a lot of time in meetings to understand my stakeholders’ questions and what they really want to know about the data before I start modeling. It also gives me a chance to see what people think of the data and areas of potential biases in the data collection or perception.
It’s also important to have at least several continuous hours of time to code and explore data, as it’s really messy. Data science is not a fast process or a magic bullet where we can shove data in a model and get great results. It’s actually really detail-oriented and sometimes fills many days with debugging code or fixing data sources.
What surprises people about what you do? Some people think data scientists are only technical resources who do a lot coding, and work alone. This cannot be further from the truth. Data scientists should be involved in a wide range of conversations to understand nuances of the business that can be captured in data. This is why we need data scientists with diverse backgrounds to join the field. Also, data scientists are constantly learning about new algorithms and different code languages. You really never stop learning in this field. Technology tools are evolving quickly, particularly in the data science marketplace. As a consultant, you are especially challenged to learn different tools that are right for each client.
What’s the best part of the job? Meeting different people and getting to understand their areas of expertise. Because of data science I have gotten to interact with cloud business planners, atmospheric scientists, economists, engineers of different disciplines, marketing — you name it. It’s really cool to learn from other disciplines and see the problems they face regularly. Learning from an industry expert about how a process works is my favorite type of interaction.