Cohort 10 Student Spotlight: Meet Lucas De Oliveira
Lucas is a student in Cohort 10 of the Master of Science in Data Science program at the University of San Francisco. He joined the MSDS program after completing his Bachelor’s degree in Economics at the University of Virginia, playing in a band in Los Angeles, and working for a tax incentive company.
Background
Evie: Hey Lucas, can you start by telling us a little bit about your background?
Lucas: I’m originally from Brazil, and I moved to Richmond, Virginia when I was young. In college, I studied economics and math with a focus on international trade — I’ve always liked systems and wanted to explore them more within the context of economics and math. I would say that I am fairly quantitatively inclined, but after college, I moved to LA to play music and I did that for two years. I was also tutoring during this time — mostly math — before I got a job at a tax incentive company, where I noticed that some of the processes could be streamlined. I had some experience with R and Stata from studying economics, and from there I started teaching myself how to program because I felt like I could make life easier.
Why Data Science?
Evie: That’s so interesting! So what made you want to pursue a career in data science?
Lucas: It took some time for me to know in what direction I wanted to push my career, but eventually, I learned about the MSDS program, and data science seemed like the solution to every problem I was up against in my job, so I thought, “I’d better go to school for this.” It’s cool having the tools to navigate really complex problems. Honestly, data science ended up being a lot more about deep learning and machine learning solutions to things, which is very different from the econ approach to quantitative analysis of attempting to explain relationships (for example, linear regression versus a Lasso regression model or a random forest model.) I like how data science here gives us the tools to handle more complex problems and focuses more on prediction and performance rather than simply trying to understand relationships that we might not even be able to understand.
Why the USF MSDS Program?
Evie: That makes a lot of sense; I think with me, coming from a psychology background, I feel the same way about the difference between those two approaches. So in particular, why did you pick USF’s MSDS program?
Lucas: I chose this program for a few reasons: the practicum, the fact that it’s only a year long, and I’ve only heard and read good things about the program. I have some friends and family friends that are tied to USF who told me to check out the program and it looked really cool to me. I knew that I wanted to go back to school for data science, but I wasn’t sure if I wanted to do grad school part time while I worked or if I should commit to doing it full time. USF seemed to be the best mixture of a really rigorous program, great faculty, very quick turnaround time, and location. I was also impressed by the fact that this program had ethical considerations around data.
Evie: Definitely! I really appreciate the different perspectives on data science that the program offers and I’m looking forward to the Ethics in Data Science course that we will be taking later in the program.
Favorite class and project?
Evie: What would you say has been your favorite class and your favorite project so far?
Lucas: So far, I think the data acquisition class has been my favorite. I feel like we learned a lot in that class really fast, and it serves as a great resource for lots of the initial tasks at hand in a data science problem, like if you need to scrape a website. I also loved learning about how the Internet works, knowing about different data formats, how to set up a server — I was really impressed by how much I learned in that class.
Evie: Same here! I loved how hands-on that class was, and despite how challenging the projects were, I always walked away from them thinking, “Wow, that’s so cool.”
Lucas: I also really like the machine learning classes. As for my favorite project, I would say the gradient descent project that we did in Introduction to Machine Learning. Even though I think we did a lot of fun projects in data acquisition, the gradient descent project definitely taught me the most for machine learning. I think getting to learn about what goes on under the hood in an iterative solution like that is really interesting and set my brain up for this module really well.
Areas you are most interested in?
Evie: What areas of data science are you most interested in?
Lucas: I think coming from my background, I’m already really comfortable with linear models. Now, I’m interested in deep learning and control systems, like having an agent or something that’s making decisions. Specifically, what I want to do as of now is apply machine learning and optimization methods to energy management and distribution, and I think there’s a lot of room for efficiency and optimization in that field.
Practicum
Evie: I think that leads well into our next question, which is, what is your practicum project?
Lucas: I’m doing my practicum at Nextracker, and I’m really excited about it. It was one of the projects that when I saw it, I knew this project was for me — it was exactly what I wanted to do. There was one added component to the project that I didn’t foresee, but this is how it’s going: Nextracker sells the tracking hardware and software that hold up an entire array of solar panels on huge solar farms. These trackers track the sun to try and maximize output. In my project, I’m in the process of optimizing those trackers by finding the right angle for the solar panels to be at at any given time. There are certain things like whether one row is casting a shadow on another row, that can drop production by 50%. If it’s cloudy outside, we want the panels to be more horizontal since most of the radiation at that point will be diffused. Right now, I’m doing manual simulations to see what angles get the best results, and later, I’ll be training a deep learning model for this optimization process and forecasting grid prices, which would involve a battery and understanding when to output into the battery versus into the grid to maximize profits. I’ll also be forecasting power production. I get to work with a lot of the management at Nextracker to incorporate what I’m doing with the work that the other engineers at the company are doing. Some of the work is very math-intensive, but they have been pretty good about explaining the concepts and the packages they use to me, and in my first few weeks, I also scripted some functions that have been really helpful for me starting out.
Personal Interests
Evie: What do you like to do to unwind and how do you spend your free time outside of the program?
Lucas: I live right by Golden Gate Park so I like taking walks with my dog and my girlfriend, and I like hanging out with my friends and going to new restaurants. I’ll still pick up my guitar from time to time as well!
