How I Became a Data Scientist

sheet-music-277277_1920

The following post was written by Chris Seal, a recent Galvanize Data Science graduate.

As I sit on a flight back to Cincinnati on this chilly October morning with my laptop open – a cup of coffee secured between my knees – I’m realizing I would be remiss if I didn’t reflect upon my three-month tenure as a student in Galvanize’s first Data Science Immersive cohort in Denver. My trip to the airport this morning seems to be a perfect microcosm for my experience in the program.

I woke up at 3:25 a.m. to catch a 6:40 a.m. flight. Since I didn’t have a car, didn’t want to pay for a taxi – and didn’t want to ask any friends to drive me to the airport at four in the morning – I ended up walking my luggage to the nearest light rail stop,1.3 miles away. For a walk that usually takes me 20-25 minutes, I gave myself 35 minutes. Plenty of time, right? Sure enough, a little more than halfway through, I looked at my phone and it was 4:08 a.m. – 12 minutes to departure! With a 54 lb. piece of luggage in one hand, a 37 pounder and a guitar in the other, and a fully-packed backpack, I started jogging as fast as I could to the light rail station. By the time I arrived at 4:18 a.m., I was drenched in sweat in 50o F weather. When I checked into the airport, I played ‘musical shoes’ to my bag to 49.5 lb., power-walked to my gate, and arrived just as my zone was being called. In other words: Everything went smoothly.


It was a hot day on July 1st, when I was driving a U-Haul cross country – my wife and I were moving from Boulder, CO back to my hometown in Cincinnati, OH / Northern Kentucky. I had applied to Galvanize several weeks beforehand, had a couple of interviews, and when I didn’t hear back, I just assumed that I didn’t get in. Somewhere between the cornstalks of Kansas, I glanced at my email and sure enough, I was accepted into Galvanize’s Data Science Immersive program in Denver! The cohort started on July 6th – five days away. I seriously considered just turning the U-Haul around.

Soon after I got settled into Cincinnati, I booked a one-way flight to Denver that got in on Sunday evening before class started on Monday. That first week of class was a whirlwind to say the least. I slept on my cousin’s couch for a couple of nights, then in my friend’s guest room, while I was trying to figure out how to find a more permanent place to stay, pay for the program, eat, sleep, and…well…learn data science. I thought about putting a tent up by the Platte River, and just staying there all summer. Upon reflection, (and if it were legal) it wouldn’t have been such a bad idea.

All that to say, it was pretty cool to be there. Some wickedly smart students, a few knowledgeable TA’s, and an instructor who seemed to be able to answer any question that was asked of him as if he had planned out his perfect response the night before. In my opinion, that Data Science classroom had a kind of ‘this is where things are happening’ aura to it.

But that first week was pretty overwhelming. I was sleeping about five-hours per night (which unfortunately didn’t improve too much as the cohort went along), and because I got accepted so late, I didn’t have time to do any of the pre-course work. Stir in the fact that our assignments were usually too long to be finished in the allotted time and that I had less recent experience in programming and data analysis than did most of my peers (I was a professional music composer, after all), and it was a recipe for feeling overwhelmed for most of the day, for most of the days.

There were times in the first couple of weeks that I briefly considered dropping out of this cohort and reapplying in the next cohort after I had more time to prepare. But instead, I made the conscious decision to stay in the program only if I worked harder than everyone else. Who knows if I actually accomplished this goal, but it was a great way to stay motivated. This was a really hard-working group, so in order to work harder than them I had to study data science at least as much as my body would allow, sometimes more.

During class-time, I worked just like everyone else did. Though, I did ask a lot of questions. Even if I had an idea of what the answer could be, I still asked the question anyway – sometimes just to hear how the instructor thought. The class isn’t free, after all, and I didn’t hold back from asking questions, even if there was a chance they might make me look dumb. My goal wasn’t to impress my classmates or even my teachers; rather, to begin a successful career in data science.

Even though I learned a lot during the day, it was the evenings and weekends when I truly compartmentalized the knowledge. I tried a lot of things outside of our daily requirements, including: writing out coding techniques by hand, writing out the key points of the lecture decks in my own words, recreating my peers’ solutions to the paired programming assignments on my computer to see if I could learn from their respective styles, and more. When I discovered that I could download instructional videos to my phone from sites like Khan Academy, Coursera, and Udacity, this was a huge bonus for me. (One of my peers, Hugh Brown, even gave me a script so I could download relevant Youtube playlists on machine learning.) My typical daily commute was 1 to 1.5 hours on the bus and/or light rail each way. Because I was often too tired to actively work on assignments, half-paying-attention to videos like these was the perfect way to gain more repetitions of the machine learning material and improve my foundational math, stats, and probability chops. I even watched videos when I walked to/from the bus or went jogging. Again, I did as much as my body would allow. My approach to learning was to get as many repetitions as possible. Even if I didn’t understand every single detail at each individual repetition, every time I was exposed to the material I would learn something new and solidify something else.

But no one else knew I was doing this, and because of that there were times when people may have underestimated me a bit. But it didn’t bother me. Generally speaking, I think it’s easy for us to let our initial impression of someone cloud our current opinion of them. I’m guilty of it too. My point is if you find yourself in a similar situation, just keep believing in yourself and working hard. It will pay off. Trust me. I know because it did for me.

When it came time for our final capstone projects, I think I surprised a lot of people. The capstone projects take place in 2.5 weeks and are intended to demonstrate our newly acquired data science expertise. Multiple people told me multiple times that my project was too hard to do in the short timeframe, and that I was trying to go too far with it. They made a good point, and I had the choice to listen to them – but I didn’t. Because I have taken on so many creative projects, I have developed a sense of how to distinguish the signal from the noise, so to speak. I was the only one that understood my vision thoroughly, so I trusted my gut and kept working.

Now was the time when the discipline, knowledge, and the skill sets I had developed through the first part of the program started to really pay off. I worked so hard. I’d often go to bed at 2 a.m., wake up before 8 a.m., promise myself I was going to go to bed early that night, only to start making progress in the evening and go to bed at 2 a.m., repeating the process. (It was like a twisted version of Groundhog Day where the main character tries to debug his code, instead of figuring out how to ‘get the girl’.) I usually need 8-9 hours of sleep a night to feel functional. To remain sane, I made it a point to sleep in as much as possible on weekends. This helped a lot. During the week, I would get exercise by doing four sets of anaerobic exercise twice a day at the gym downstairs. This would take 10 minutes or so at a time, so I wouldn’t quite break a sweat. Very efficient. I also did about an hour of walking a day on my commute.

I stopped watching instructional videos during this time, and instead, just let my mind wander while in transit. More often than not during these times, some great ideas for my project would pop into my head. From what I understand, this is thediffuse mode of thinking. If you’re not familiar with the term, I’d suggest you look it up and use it to your benefit.

On Monday, three days before our Capstone Showcase, I was working on the 2.0 version of my script. I had written it over the weekend based on version 1.0, and was trying to debug it all (I wouldn’t necessarily recommend this approach). I came close to finishing it, but ran out of time. At 11pm on Monday night, I finally gave in and started working on my presentation. I spent all day Tuesday doing data visualizations, Wednesday daytime preparing my presentation deck, and Wednesday late-evening re-creating my resume. On Thursday morning, I awoke in typical sleep-deprived fashion, buttoned-up my favorite shirt, tied my hipster skinny tie, and headed out the door. I mentally rehearsed my speech on the light rail trip to campus. I think I made my last edit to my deck about five minutes before the event started. On too little sleep and a wholly inadequate amount of caffeine, I presented my capstone project, conducted micro-interviews with ten companies, and networked over lunch. In other words: Everything went smoothly.


It may be tempting to see the high median salaries in data science, and think joining a program like Galvanize’s Data Science Immersive program will get you there. If you have the Galvanize credit by your name and practice the interview questions, you may be able to fool some people. If you take a couple of Data Science MOOC’s, you may claim on your resume that you know how to run certain algorithms, even though your level of expertise may be quite superficial. In my opinion, the only way to truly become a data scientist is simply to put in the hours.

But that’s not the biggest lesson I learned. Throughout the program, people thought of me as a professional music composer who was learning data science. The assumption was that I had always been a good musician. In reality, up until my sophomore year of college, I couldn’t read sheet music. I was in music theory classes with peers who had played an instrument and studied music theory since they were in their early single digits of age – and they had the ears to show it! I was just a rock guitarist who occasionally jammed out to popular alternative rock tunes in high school. Quite a difference. There are parallels here to how I felt at times in Galvanize’s Data Science program. I was surrounded by PhD students in analytical fields, software developers, etc. At times, I felt out of place. But then I remembered a past conversation I had with a music professor. It was right after I had performed a Filipino guitar+vocal duo with my now wife in front of a group of incredibly talented pianists and vocalists. I was more of a composer than a guitarist at the time, so my chops were not even close to on par with theirs. I told the professor afterwards that I felt intimidated performing in front of such a talented group – and I meant every word of that. He simply responded: You’re one of them. (And you probably are, too.)

My plane is landing now, so I should probably get going. Data science is such a broad field and there’s so much to learn, so many skills to improve upon, so many problems to solve. Galvanize helped me get started, and it’s up to me to keep it up. Someone told me it’s like having a baby: after a certain number of months, you plop it out, and then, it’s up to you to spend the next few decades investing in it. Perhaps, that’s why I find this field so interesting.

Want more data science tutorials and content? Subscribe to our data science newsletter.

Video Transcript

Ryan: Data science is a relatively new field. People always say it’s you know it’s either statistics, or machine learning, or computer science, and these have been around for a long time. But the thing that’s special about it is how these concepts come together in industry. And largely it’s about pulling out insights from your data and being able to build better products. Make better decisions, and largely build a better company through the use of this huge torrent of data.

Katie: People in our program are coming from various backgrounds. That include engineering, analytics, and the quantitative sciences.

Man 1: I’ve been a professional poker player since graduating college.

Aznah: I was a PhD student in planetary science.

Man 2: I was a professor before coming here.

Conner: I worked on Barack Obama’s reelection campaign.

Man 3: Working as a software developer I had a very renew-ability seller.

Dan: What really drove me to poker in the first place was that it was a problem to solve.

Conner: I was using Excel and other tools like that, but I knew there was so much more out there. And I wanted to become a part of.

Aznah: I was working with lots of data sets, but wanted to get into some real world problems.

Conner: The days here are non-stop. We start in the morning with a mini quiz.

Aznah: The afternoon pair exercises are sort of an exercise in collaboration as well as working through a problem.

Conner: Steve Twa showed the coolest presentation where he was doing a live demo of playing songs on his violin, and putting out what song he was actually playing. And showing how machine learning could do that audio processing. So I was blown away by that. I actually really lucked out they connected me with Steve and he became my mentor. It was incredible to see how quickly we kind of all came together and were able to learn a ton of material in such a small amount of time. I could reach out to someone. Show someone my code and have that instant feedback.

Aznah: Everybody is very supportive here. We’ve formed some really close friendships I think.

Ryan: It’s a network of hiring partners, and alumni, and data scientists throughout the field who really come together. And are helping shape and define what data science is, and what it means for the future.

Aznah: My name is Aznah Azhari.

Dan: I’m Dan Morris.

Conner: My name is Conner Pinkston and I’m a data scientist.