Summary

In this second post of a three-part series, I reflect on my Master of Science in Systems Science degree. Here, I take note of what I did well and what I didn’t do as well over the course of my degree.

---

What I did well:

• Enrolling with a clear goal

I knew before starting the program that I wanted to be a data scientist: this has guided which courses I take, which projects I work on, and how I spend my time. I’ve applied systems science techniques and methodologies to concrete problems I’m interested in (for example, using a genetic algorithm to optimize routing as a term project for Artificial Life, and a successful term project model in Discrete Event Simulation to prescribe bike allocations to Biketown stations that I deployed at work). It’s easy to get lost in abstraction of systems ideas but having a clear goal of where I wanted to go and how I was going to get there was useful.

• Finding and/or creating my opportunities

I got my first data job while catching up with my mentor from undergrad. She mentioned that she had some datasets but didn’t have the time analyze them, so I suggested that I could help her find insights in them. It wasn’t a glamorous job, but it was my foot in the door. I learned a lot, helped the Rec Center on campus, and set myself up for my second job, at Biketown.

My job at Biketown came through my graduate certificate in Business Intelligence & Analytics. There was a hotly contested opportunity to work with Biketown and my group won the coin flip. We spent the term working on a project which i took the lead on. After the project, I reached out and talked my way into being hired on.

• Taking classes from other departments

The flexibility in coursework I could apply to my degree appealed to me. In addition to the Masters, I enrolled in a Business Intelligence and Analytics graduate certificate where I took classes such as Operations Research, Data Mining with R, and Data Warehousing. I was able to stitch together a data science curriculum with those classes and Multivariate Data Analysis and count most of them to my degree requirements.

• Prioritizing restedness and decompressing

It’s hard to think or learn new material when tired, and it’s even harder to write code that works. Keeping myself rested let me absorb information faster and think clearer, which meant I didn’t have to spend extra time studying because I was tired (which is a vicious cycle you don’t want to get into). I also made sure to budget time for socializing with friends – grabbing a beer and decompressing with them was important for maintaining mental sanity and avoiding burnout.

• Setting project scope and getting started early

I tried to get concrete and actionable project ideas early in the term, which allowed me to start thinking in terms of deliverables and what I needed to do to get to them. Code written under pressure is never as good, and you never know how long it will take to complete. When others were stressing about their projects near the end of the term, I’d be putting on the finishing touches. Starting early kept me stress-free and let me produce good work.

What I could have done better:

• Taking the theory courses before their application

While I appreciate the elegance of theory,I prefer tools and frameworks I can use to solve problems with. Most of the systems science classes I signed up for my first year were on applications of systems ideas (e.g., how to model different kinds of systems) before I really understood the systems themselves and how they behaved. Most of my systems science classes this year were more theoretical, such as systems philosophy or using information theory as theory to guide data analysis. Having the theoretical foundation would have helped me understand systems science better overall instead of trying to backfill knowledge – and I wouldn’t have needed to cram systems theory before graduating!

• Creating a community with data people

I know a lot of friendly faces from people I’ve seen at R and Python meetups, DAMA events, or other similar gatherings. They also know me as a friendly face, but we don’t always know each others’ names or what kinds of projects we’re working on. I should have done better over the last two years at getting to know them better, reaching out to ask to grab a coffee and chat. This is luckily something I still can and will do.

• Explaining what systems science is

Systems science is fundamentally difficult to explain because its scope and purpose are unique to each individual. It’s such a broad topic that I think its ideas are relevant to literally everyone. I’ve bounced around between trying to find a relatable answer by using what I know about the person who asked me the question to make my answer relatable, or using a canned answer; neither of these is fully satisfactory. I’m going to need to find a good canned answer that I can augment with information relevant to the person, even though this cold-reading is difficult to do.