How I, a Journalist, Survived Grad School and Earned a Data Science M.S.

Steve Jobs once said coding is a liberal art; I put Jobs’ interpretation to the ultimate test, with no computer science or mathematical background.

In 2019, with both demand and salary for data scientists increasing exponentially, I decided, barely two years out of undergrad, to enroll in a master’s degree program with the goal of becoming a full-fledged data scientist in 15 months.

It was one of the most rigorous academic challenges I’ve ever undertaken.

The purpose of sharing my story isn’t for self-congratulation or to suggest that my path should be emulated. In fact, I would advise anyone considering entering the data industry to acquire at the very least a foundational knowledge of computer science, database architecture and to take a statistics course or two.

Instead, my goal is to demystify the barrier for entry into both graduate school and the data science field overall. In a field in which nearly 50% of data scientists hold a phD and positions often list a master’s degree as the bare minimum requirement, it can be intimidating to even think about applying to school, let alone pursuing a data-driven career.

A motivated individual transitioning to data science can succeed by leveraging prior domain knowledge, engaging in structured self-learning time and completing independent projects to build a robust portfolio.

Below, I’ll share how eventually, through brute force, I made it to graduation day and how you, the interdisciplinary data science hopeful, can do the same.

I Put Down the Books and Read the Docs

Although I recognize the depth and accessibility of books, when it comes to programming, I didn’t learn that way. Don’t get me wrong, I still did a lot of reading. I read documentation and, more importantly, accompanying code snippets. While Stack Overflow offers case-by-case fixes, there is truly no substitute for the comprehensive learning that can be accomplished simply by reading documentation. Instead of Googling how to create a pivot table in Pandas, you’ll save much more time by doing the work and reading the docs. Even post-graduation, prior to beginning my current role, I spent two weeks reading BigQuery documentation and learning, function by function.

Like books, the more documentation you read, the more you can recognize what constitutes a quality and intuitive guide to a programming language, library or application. As a data engineer, I can tell you, there is a spectrum of clarity when it comes to API documentation, in particular. My preexisting reading comprehension skills helped me retain information and implement concepts.

If you’re an English, History, or Journalism major, you’d be surprised how critical thinking skills translate to interpreting and writing code. Assigned reading in school prepares you for how much technical reading you’ll do working as a data analyst, data scientist or data engineer. Additionally, coming from a background in which you’ve learned or leveraged communication skills can prove to be an advantage when it comes to commenting on code, collaborating with peers or presenting yourself to future connections and employers.

I Pursued Datasets within My Area of Expertise

If you’re a current data science or data analytics student you can likely attest to how mundane Kaggle datasets or textbook data sources can be. There are only so many ways you can predict Titanic survival rates or to classify flower species. When it came time to choose a semester project, I did my best to find datasets that aligned with my prior knowledge. I found that picking data from an industry I was passionate about provided me with both credibility and inspiration. Having domain knowledge also enabled me to better contextualize and present my findings, leading to a better presentation and a more fulfilling learning experience. In school, coming from journalism, I prioritized the why of my data work, a skill you too can learn by reading my previous story.

One of my favorite data science projects I worked on was leveraging various classification methods to build a fake news detection model in Python. Looking back, the approach was fairly simplistic and not incredibly novel, but I was able to thoroughly articulate the problem such a model would attempt to address. The project was also the first time I expanded beyond the datasets native to languages like Python and R. Post-graduation, I continued to seek out and create novel datasets by scraping web data and accessing APIs.

One of the most important lessons I learned is that it’s much easier to finish a passion project than it is to slog through an assignment.

Therefore, whenever you have an opportunity to pick your own projects, make them as interesting and applicable to your professional skillset as possible.

I Took (Even More) Classes

Even though I took two to three courses per semester, I identified gaps in the material presented and my own understanding fairly quickly. In my second semester I discovered platforms that offered free or heavily discounted courses on hyper specific topics like advanced Tableau functions. If you’re entering the data world without a thorough understanding of computer science or data structures, I suggest you check out the following resources:

• edX
• FreeCodeCamp
• W3 Schools
• Udemy
• DataCamp

Data science is a massive field and even university courses will have gaps that need to be addressed. As I mentioned, I’m a huge proponent of self-learning. The ability to teach yourself a concept in a limited time frame will be invaluable in an ever-changing industry. In a data-oriented or any other technological role, not only will you be expected to learn a company’s way of using technical tools, but you also may be expected to participate in continued learning programs or work toward certifications, like those provided by Google Cloud. You might as well start learning how to teach yourself while still in school.

I Asked for Help

In addition to meeting with my professors, I received assistance from two amazing tutors for Python. If, like me, you entered school with little technical background, individualized help can reinforce concepts and help build confidence. While taking a ten-week data mining course later described by a dean as “Phd-level”, I met with a tutor twice a week to translate pseudo code and simply stay afloat. It was an intense, stressful summer and the most satisfying ‘B’ I’ve ever earned. However, being able to discuss and work through assignments with someone other than a professor helps alleviate stress and can make the topics and concepts less intimidating. Another benefit of tutoring is that, for programming, tutoring mirrors Agile’s pair programming model. It’s likely that, as a new graduate, you’ll report to a senior data analyst, lead data scientist or senior data engineer. Becoming comfortable enough to code and, more importantly, make mistakes, in front of supervisors is a skill in and of itself. Working with a tutor can help you embrace the idea of someone looking over your shoulder while you code in a way that is constructive and productive.

I Practiced

I learned, several semesters in, that I was never going to become proficient at programming and data work simply by doing school assignments. Learning coding is like learning an instrument. You’re not going to be proficient if you just show up to practice.

A good litmus test for determining if a data career is right for you is how much work you put in outside of school or work; genuine curiosity and a willingness to learn can provide motivation to push through tough concepts.

Personally, even with a 9–5 data engineer job, there are still aspects of the field I find fascinating. For instance, I’m interested in how I can use data engineering to improve or automate my life, like how I created a personal finance tracking pipeline.

Practice doesn’t just hone skills. It builds confidence. In my experience, you’ll be far more likely to solve problems by losing the fear associated with making an error or generating the wrong output. Depending on your trajectory and language of choice, I highly recommend HackerRank, a site that provides plenty of practice opportunities; bonus: Employers use the site for technical interviews.

Don’t Follow My Path — But Consider My Advice

Although I was able to transfer from a heavy liberal arts background to data engineering, my path was very atypical and, frankly, in the beginning, very difficult. However, I believe that teams and organizations benefit from technical workers hailing from all kinds of backgrounds. Even though you may be a data scientist, you’ll likely work in or specialize in a certain industry. Coming from an interdisciplinary background can help you identify which industries align with your previous experience, domain knowledge and passion. If you’re considering exploring a data science or data science-adjacent path, please know that there are plenty of resources available to empower your own learning — even before you apply to that M.S. program.

Create a job-worthy data portfolio. Learn how with my free project guide.