The data team at AdvancedMD just had its first birthday a few months ago. Last year, the leadership decided it was time to invest in their data strategy by creating a data team with two analysts, a data scientist and a data engineer. Since then, they have revolutionized how data is accessed and used throughout the company. I spoke with Dallin Homer, the Business Analytics and Insights Manager over the data team, about how data science is done at AdvancedMD and any advice he had for those getting started in the field.
What’s a day in the life of a data scientist here?
“We haven’t figured out what a ‘traditional day in the life of a data scientist’ is yet, but we’re closer to it now that we have a data engineer. Before that, our data scientist was doing most the engineering and setup for our platform that we have today.” Now that their data engineer is handling most of the data pipelines, the data scientist is working on more predictive models for client retention and revenue anticipation using machine learning.
“By way of tools used, we use SQL, Python and some R for doing data mining from our CRM and databases and rely heavily on pandas for preparing the data for what we’ll need downstream.” For predictive models they leverage the sci-kit learn library especially XGBoost for classification models.
What are the backgrounds of the people on your team?
“We have a mix of bachelor degrees in Economics, Business and Computer Science. A few of us are currently persuing master’s degrees in analytical fields. Most of us have spent hours on the side learning and improving our skills.” Their data scientist completed a data science bootcamp in NYC and had worked previously as a data engineer and scientist before joining the team.
What are skills or attributes you look for in new data scientist hires?
“A real passion for data that can be seen through examples of personal experience – either on the job or on the side.” Dallin says examples where applicants can show how they “[got] familiar with data … and then did the data preparation and engineering to explore and solve a business question” is also a good sign for an applicant. He expects them to be able to communicate how they worked through their projects so he can see how they think through problems.
Technical skills are of course important. A junior data scientist should be familiar with SQL, command-line skills, pandas and Python fluency and have experience with data modeling and evaluation. “They should also know how to answer the question ‘How do you know if your model is performing well?’”
What would you say to someone just starting in data analytics?
“Find something that interests you, whether it’s sports or weather or whatever, and think about how you could present it visually in a compelling and simple way.” This is a great way, says Dallin, to start thinking about how data analytics could apply at a business.
He recommends connecting with people in this space and learn the tools that are being used to solve your problem and start learning those. “Be inquisitive. Always try to understand the data before you set off to model it. Understanding the context of the data is a huge asset.”
“If you’re not sure or don’t know what interests you, try looking around in your community for problems that you could tackle. Ask a family member or a friend if they have any projects. Just start going through the motions. Never underestimate the power of the side-project on your career.”
My entire career has been powered by the synchronicity that comes from side projects.— harper 🤯 (@harper) April 24, 2016