Making the Switch from Civil / Environmental Engineering to Data Science
I’d like to start with a disclaimer that this is my personal experience entering the data science field and may not reflect the path the majority of data scientists take. Data scientists come from all different backgrounds and everyone’s journey is unique.
Since I’ve started working as a data scientist, I’ve received inquiries from fellow civil and environmental engineers about how I made the switch. This post summarizes my responses to the questions I am typically asked:
- When did you decide to make the switch to data science?
- How did you get started learning data science?
- How are you liking it so far?
- Were you able to leverage your civil engineering background to help you get an edge?
- Is getting a P.E. license worth it?
- Did you have enough statistics and programming background to feel comfortable transitioning into data science?
- What are some other data science resources?
Edit 2/2021: I’ve also frequently been asked about how to frame civil engineering job experience when applying to data science positions. Here’s an article I wrote to answer that question.
When did you decide to make the switch to data science?
Like many data scientists of today, data science was not a degree option when I was in college. I took one computer science class to complete my engineering degree, and it was my least favorite undergrad class. Primarily because the teacher was awful and C++ is probably the least useful coding language for civil engineering (even Fortran would be have been more useful!). I never in a million years thought I would consider coding as a profession. My brother was the computer scientist — he grew up building computers and having LAN parties. I, on the other hand, dreamed of a career where I gathered water samples in the mountains and wrote detailed analyses about the water quality and would eventually save the world from climate change. It turns out very few people in the civil engineering industry get paid to research the water quality of mountain streams.
After I graduated, I diligently put my bachelors and masters degrees to use by working as an environmental consultant. And I LOVED it … for about two years until I got burned out from working too many hours. I lost all my motivation to advance my career in environmental consulting. I kept getting more and more frustrated by how inefficient, bureaucratic, and archaic the industry was (don’t get me wrong, I still have mad respect for civil engineers).
Then I was conveniently offered a new position at a tiny start-up company to guide a team of software developers to build out software related to environmental compliance. I took the position as a product manager and dove into the Tech industry. I even took a Product Management night class at General Assembly to learn as much as possible as quickly as possible. After 6-months working at the tiny start-up, we burned through our funding and I was laid off. This turned my world upside down and forced me to make a decision about which career I wanted to pursue. After applying to a few product management positions, I realized that I didn’t really like product management. I wanted something more technical. I wanted to feel like I was contributing to building something innovative. So I started learning to code.
I quickly realized that I did not want to be a programmer. I was not a person who could spend all day coding; I needed more variety in my tasks. That’s when I stumbled across data science back in 2015. After looking more into it, I realized that I had already used data science to complete my masters’ thesis research in grad school, and I really enjoyed it! But I had a lot to learn before starting a career as a data scientist.
How did you get started learning data science?
Once I realized that I wanted to pursue data science as a career, I kicked off my training by learning to code. I built a few rough websites from scratch using an Udemy course as a guide (I don’t recommend the one I completed because it’s very out of date now). Building a website gives you an idea of how software applications read from and write to databases, which is very important for data science. I also started to learn SQL and Python via Codecademy to get an intro to the languages (I had already used R in graduate school). Eventually, I started the Udacity Data Analyst Nanodegree to solidify my Python, SQL, and intro to machine learning skills. I took a year to complete the Nanodegree (I did the class work in the mornings before work and on the weekends). Udacity now has a Data Science Nanodegree, which is probably much more applicable than the Data Analyst one.
After the Nanodegree, I applied for a few jobs and quickly realized that I didn’t have the right skills yet. I needed to build out a portfolio of projects to prove I could complete useful tasks as a data scientist. I joined Kaggle and participated in a couple machine learning competitions, but primarily Kaggle was a great resource to understand how other data scientists were approaching data analysis and machine learning in R and Python. I also went to a conference to try and network to find a job, but data scientists are not the most social bunch.
What I did get out of the conference is an understanding of all the latest tech that companies need data scientists for (chat bots, text mining, machine learning, etc), and there were super helpful tutorials! I highly recommend attending a conference with workshops although it’s incredibly expensive if you don’t have an employer cover it.
At this point I was hitting a roadblock: I was starting to forget things I’d learned the year before because I could only study before work and on the weekends. So I decided to quit my full-time job to build out a portfolio of projects and apply to jobs full-time. I expected to be unemployed for at least 6-months.
About 2 weeks into my (f)unemployment, the software company I previously worked at reached out to see if I’d be interested in building out a risk model for their software. They just received a big round of seed funding. Of course I said yes - what an opportunity! So that is the story of how I got to my first data science job (2 years after I decided to embark on the career change).
How are you liking it so far?
I love it! I think it’s challenging, requires creative solutions, and I’m building something innovative (everything I thought civil engineering would be…)!
That said, I’m enjoying the product I’m working on, I have a diverse range of tasks at a small start-up company, I find it rewarding, and I have great mentorship from a CTO when I get stuck. Without each of those pieces, I would probably like it less. I’ve spoken to others who switched to something like advertising on social media… they do not like it because they don’t find it rewarding. They’d rather do something more related to the civil/environmental industry that they started in.
Also, if you’re just starting out in the field, you want to be sure to set yourself up for success with great mentorship at your first position. I had multiple interviews where my takeaway was that the company needed to hire a data engineer, not a data scientist. If I had taken a job like that, I may have set myself up to fail or at least not meet the company’s expectations.
Were you able to leverage your civil engineering background to help you get an edge?
I think the only way I was able to get a job as a data scientist is by leveraging my civil engineering background. I can’t state it enough: it’s a very competitive industry to break into. While data science may be the most in-demand position of 2018, companies are looking for data scientists with proven experience. Hiring someone great is already difficult, so it’s a big (expensive) gamble to hire a data scientist who has no proven coding experience.
Is getting a P.E. license worth it?
I’m really glad I got my P.E. license. It makes it easier to return to the civil engineering field if I need to. Plus, I actually use it at my current position to review technical reports and design calculations. That’s just me personally, I think there’s no right or wrong answer to this question — it just depends on the person.
Did you have enough statistics and programming background to feel comfortable transitioning into data science?
I was actually applying data science as part of my masters’ thesis research in grad school, but it was called biostatistics. I took a lot of statistics classes as part of my MS degree, which helped immensely when breaking into data science. I had some coding experience in R, but not in Python or SQL. I don’t think it’s a barrier not having either a coding or statistics background to start with, but it’s definitely a challenge and it will take longer to learn the field.
What are some other data science resources?
Masters Degree. This is probably the best way to go if you want to get hired quickly. I didn’t choose this because I already had an MS degree that I was still paying off.
Bootcamps. There are quite a few bootcamps out there. Supposedly, certain companies hire from certain bootcamps, so it’s best to pick your bootcamp based on the company you’d ideally work at. I personally didn’t choose the bootcamp route because: (1) I felt that I had already learned enough on my own, (2) I didn’t want to quit my full-time job, and (3) it only gets you part of the way to a job (expect to be unemployed 4–6 months post-bootcamp). However, it is a great resource for developing a network and gets you up to speed much quicker than learning on your own part-time. I have specifically heard good things about these bootcamps:
Websites not previously mentioned. The internet is a great resource — there is an overwhelming amount of free data science classes and materials.
