Advice on Transitioning Industries - Economic Research to Data Science
Since I graduated from UCSD in June 2019, I have worked as a research analyst at the Federal Reserve Bank of Philadelphia. I assisted economists with economic research projects and compiled data on the U.S. economy to prepare monetary policy presentations. This fall, I began a new job - as a senior research analyst at Accenture. While my job title remains (nearly) the same, I am now working on Accenture’s global data science team and applying research to the data analytics and AI space. For me, the industry transition was unexpected (a story for another article) and full of uncertainties - how could I prove to potential employers that my skills were transferable and that I was a qualified candidate for a field where I had no prior direct experience? Making an industry transition in your career involves taking on many unknowns - I’d like to share some advice on how I personally addressed these challenges.
For complete context, I spent about eight weeks seriously recruiting (from the end of August until the end of October). I applied to 87 jobs, got 9 interviews, and ended up with two offers (my other offer was for a Data Scientist role at a smaller company). I may have ended up with several other offers but began declining further interviews once these two offers came in due to time constraints. I applied and interviewed primarily with companies in tech, finance, and consulting, and only for positions that were either remote or based in NYC. Without further ado, here are the ways I managed the transition.
Stay organized.
Keep track of job-specific questions and information: I recommend using either a spreadsheet (see example here) or building your own job applications tracker in an app like Notion. I also used Google Drive to keep track of my progress on applications by creating a separate folder for each job I began interviewing with to jot down notes from my discussions with recruiters/interviewers, list questions I wanted to ask, and save resources specifically relevant to that job. While it may seem silly to be creating so many folders at first, by the end of the application process you’ll be glad you kept track of everything from the beginning. You end up writing so many different versions of cover letters and getting asked variations of the same questions, so having an easy-to-find record of each application and questions asked/answers given will save you a lot of time from sorting through everything later.
Create a centralized document: The simple but tedious work of just filling out each job application with your background information, work experience, and all the other questions asked can be time-consuming. Since it’s not fun and you’re likely applying to dozens, if not hundreds, of jobs, starting early lets you space out the mind-numbing administrative work and leaves you ample time to spend on the much more important interview prep. You don’t want to be wasting a ton of time on one application when you suddenly realize you have to track down long-forgotten experiences or personal history technicalities. This is also where keeping track of all your responses in previous applications will save you massive amounts of time for subsequent ones. While each application has its own variations of requirements, there’s usually enough overlap that you can easily take responses and re-work them. This will save you a ton of time and headache. I recommend saving the PDF of your completed application from each job and keeping a Word/Google Doc that compiles all questions/prompts and your responses. Doing this allowed me to easily Ctrl+F for prompts or keywords to find specific questions in the growing haystack (that document is now over 40 pages for me) and to copy and paste responses without formatting difficulties. Make it easy for future you, and don’t start every application from scratch.
2. Practice, practice, practice.
You’ll probably see this in every piece of application advice, but it bears emphasizing: practice answering behavioral, technical, and company-specific interview questions as much as possible. Practice on your own, with data science professionals, and with anyone who can spare 20 minutes so you can attempt your interview responses with another person. And continually update/revise/redo your answers based on the feedback you receive and what feels right to you. It’s important to be able to explain basic data science concepts, thoroughly explain any previous experiences listed on your resume, and have polished answers for routine behavioral questions. You can find tons of resources online for practice data science questions or even examples of typical questions asked at specific companies. I revised each of my common interview question responses at least a dozen times, and the final versions looked very different from what I first came up with. This is where starting early can again make a big difference - it will give you more time for practicing and revising and practicing again. Once you send job applications in, your interview answers are basically the only part of your application you can still change, so you should be spending as much time as possible refining your answers and tightening things up.
3. Utilize the many, many data science resources out there.
The main topics you want to prepare for, outside of the coding discussed in the next section, are probability and statistics. These are the foundations of data science and knowledge of them is essential to conducting accurate and insightful analysis. If you aren’t quizzed on probability and statistics topics by interviewers directly, they will likely try testing your knowledge on these through related questions on your work experience or project hypotheticals. Here are some resources for preparing for the probability/statistics questions and for brushing up on your knowledge of these fields. Overall, I would recommend focusing on having a strong grasp on the basics (how to calculate distributional statistics of interest, what does a linear regression do versus a logit, what does the p-value represent) rather than tackling the more advanced machine learning/AI models. However, if you do want to prepare for advanced topics, here are some resources on those concepts.
4. Prepare for technical assessments and case studies.
My prediction is that, if you are taking recruiting seriously and spending time practicing, you will begin breezing through the first round interviews after the first couple. The behavioral questions stay the same and aren’t exactly intellectually challenging, at least in the initial rounds. Where more candidates often stumble in the process is the technical assessment step. In data science interviews, this can take the form of a take-home coding assessment, a data science case study, or (rarely, in my experience) a live coding assessment. In my many interviews, I often had a mix of a take-home assignment with 24-72 hours to complete as well as some technical coding questions mixed into the later interviews. Companies here are trying to assess how much you really know about the programs you list on your resume and how effective your coding solutions might be. The best advice to prepare for these is in the above two tips - practice as much as possible and utilize online resources to improve. Here are some resources for practicing for these assessments.
However, there are some pointers here for how to stand out besides sending in the best code possible. Following your submission of an assessment or conducting a case study, you will be asked for your thoughts on what you did well, what you could have done better, and other self-assessment questions. It is critical to show some self-awareness here - this is your opportunity to explain any shortcomings in your responses, acknowledge weaknesses while emphasizing strengths, and justify your methodology. Hiring managers care just as much about why you used a certain method as for the exact method used. I often got better feedback from clearly justifying my use of a simple logit regression than stumbling over breaking down a random forest model.
5. Seek out mentors and repeatedly ask for advice and feedback.
I went to mentors, peers, coworkers, friends not at all in data science themselves, and even strangers whose advice and emails I found online for help. It doesn’t cost much to simply send an email with some questions, and the returns to this can be immense. I’ve had complete strangers offer to look over my resume and cover letter, be willing to do interview prep or even refer me to positions at their company, patiently answer my long list of questions about interviewing, and much more. While you should always be respectful of others' time, it doesn’t hurt to ask. Keeping up these connections can also be useful for expanding your professional network and continued learning past the interview phase. Some great places to start for finding mentors are Menti’s Career Exploration and websites like ADPList. However, simply looking people up on LinkedIn with job titles you’re interested in (or working at companies of interest) and messaging them can suffice.
6. Read others’ advice - but don’t take it too seriously.
There are a ton of helpful guides, tips, and websites out there full of advice on how to prepare and apply for data science roles, as I have linked many above. But don’t waste too much time obsessing over online recommendations as I did. There are diminishing returns to reading every single piece of advice online, and at some point, it’s healthier to just send out applications and do interviews. Personal experience is the best teacher. Also, each person has idiosyncratic beliefs and preferences (some say to write creative cover letters, some say don’t even bother with them, for example), including what’s written in this guide. Take everything you read with a grain of salt, and remember the best advice will come from those who know you best.
This post mostly focused on interviewing for data science positions. If you are looking for resources focused more on learning data science concepts or how to break into the field more generally, I recommend checking out my map on Diving Into Data Science! If you’d like to talk more about the data science field or share your experiences with me, please shoot me an email at troded24@gmail.com.