How I Became a Data Scientist

As a data scientist with a PhD in economics, I always get a lot of questions about why I chose to be a data scientist, and how was my journey like. I think it is a great opportunity to share my experience here, especially the things I did right and wrong, to help anyone who is interested in data science but comes from a different background.

Why Econ PhD

Back in 2015, I was thrilled to receive a PhD offer in economics and started my graduate study in the US. I was once an exchange student in the US during undergrad and was really impressed by the education quality here thus decided to continue my graduate study here. I didn’t think too much about the choices of majors. I majored in economics in undergrad, and I was fascinated by the ways economists interpret the rules guarding the world. Continuing in Econ was a natural choice at the time. When applying to the programs, I expressed my passion for teaching, and I was determined to be in academia as my dream was to be called a professor. I knew it would be a hard journey but was ready for the challenges.

Why not academia

As I came to the PhD program directly as an undergrad, I had vague and probably wrong impressions about what is research. I was inspired by a number of devoted researchers in different fields who are determined to extend the knowledge frontiers in certain areas, but soon discovered the difficulties of dealing with the ambiguities needed when constantly working on things without the right answers. I really had a hard time coming up with research ideas and implementing them with trustworthy methodologies. Gradually, I realized I want to work on projects that are fast-paced, where I can receive feedback and see my impacts at the early stage before I go on the rabbit hole of "revising and improving". Thus I set my sight outside of academia and started to look for potential opportunities in the industry.

Why data science

There are many potential positions for economic PhDs in the industry, as the questions we solve and the skills we develop in school align with what companies need in a lot of fields. My research background is in empirical studies, which is problem-solving using data-driven approaches. Facing a problem, how to get relevant data, how to clean data, and how to drive insights from the data to answer the question, were what I practice every day while completing my thesis. I came across Python when I was working on the data analysis process for one chapter and took online classes to get to know more about Python and get introduced to data science. That’s when my interest in data science emerged and kept growing after I found my skill sets already partially matched what needed to be a data scientist.

How did I become a data scientist

Taking online classes in Python, machine learning basics, and other data science techniques were the first steps. Utilizing Python to complete the data analysis processes in my thesis really gave me a head start in conquering this language. Then I got an idea of using machine learning algorithms in one of my thesis chapter, and implemented it with two other colleagues sharing the same interests. With this background, I applied and got accepted into an eight-week data science Bootcamp with a scholarship, and that’s when I really got my hands dirty in various projects and took a racket in developing relative skills. The Bootcamp also offered career coaching, job-searching aids, and alumni panels, which helped me understand more about what data scientists do daily, and enhance my determination of applying for data scientist positions. For more details about the Bootcamp experience, you can read the article here:

How to Benefit From Attending a Data Science Bootcamp?

As someone who benefited from working on hands-on projects while ramping up data science skills, I want to give back and I am developing a data science course with hands on projects in time series forecasting, for those who are interested in learning time series analysis and getting hands dirty with the best practices in the industry. Unfortunately, the course is only in Chinese for now, but I will have a series of blog posts coming up for tutorials and interesting practices related to this course. DM me here or on LinkedIn if you are interested in attending the course.

Applying for data scientist positions

For the ones who have followed me for a while, you might know my job hunting journey was not an easy one. It’s a seven-month-long process with a lot of rejections along the way. It’s a constant challenge to self-confidence and made me doubt myself multiple times on whether I have made the right choice. As I look back, I am so grateful to have not given up easily and kept putting myself together for the next interview. Finally, it paid off. I have also written an article to share the whole experience, you can check it out here:

How to Not Feel Like "Crap" Facing Rejections from Data Science Interviews

What I really benefited from the long job searching process, is not only the valuable interview experiences, but also all these articles I have written summarizing my preparation for each interview, and many found also useful in preparing for their own interviews:

Data __ Science Interview Preparation

My journey as a data scientist in two companies

Time flies, now I have worked as a data scientist for over one year and a half. I have switched companies in the middle, and maybe in the future I will share more details about how to apply and prepare for interviews while employed. Here I want to focus on what I do as a data scientist.

Basically what I do, in either company, aligns with my expectations. As a PhD, I might take on more responsibilities in the researching and modeling part, I have the trust of my coworkers and managers in exploring and applying new methodologies. I have also always worked on end-to-end projects, meaning that I take charge of the whole process of a project from collecting data to delivering the final product. My first job is project-based, where my coworkers and I got to work on different projects solving different problems. My current position has a fixed goal, which is making time-series predictions. While the goal is the same, as to improve forecasting accuracy, we need to keep adjusting our models and methodologies to cope with the ever-changing macroeconomic situations. Both positions are challenging but rewarding, where I get to see how my work is put into production, or how my forecasts are adopted to the upstream. The feedback, no matter negative or positive, shows the impact of my work, and that’s what motivates me to continuous improvements.

Data science is a challenging position that requires continuous learning since the world is changing rapidly, and the techniques are changing even faster. It has so many potential use cases and it has been shaping the industry for all the companies to make data-driven decisions. The more time I have in this role, the more I have realized data science is not only about fancy data analysis, advanced modeling, or complicated coding, as I assumed, but more about collaborations with different teams, revealing and communicating insights, and data storytelling. In the next article, I am discussing what makes a good data scientist and seven principles I follow based on my experience:

Seven Principles I Follow To Be a Better Data Scientist

To summarize, my journey from Econ PhD to data science contains a lot of happiness, challenges, rejections, fulfillments, and so on. I am glad to work in a position I have both passion and skills. Hopefully, this article gives a reference point to those who are in a career or any life intersection, and I will be thrilled if it somehow inspires anyone. Thank you for reading!

Lastly: