How do I become a data scientist without a PhD?

I am a data scientist with a Bachelor of Science in Information Systems. Honestly, I never planned to become a data scientist, and if I knew how much work it took, I might never venture on this road.

Data science is the survival of the crazy and obsessive. Getting a postgraduate education from reputable institutions is one way to demonstrate that obsession. Standing in line at Disneyland, the supposedly happiest place in the world and thinking about feature creation and hyperparameter tuning of boosted tree models is another way (which was what I did and angered my SO).

So how does one become a data scientist? I have formulas for that

Start with the 66 mini hell formula:

• From now on, spend 60 minutes every day on data science. Start with breadth topics MOOCs from Coursera, edX and other online resources. That covers the basics of machine learning and statistics. Don't be embarrassed if you don't understand anything. Ask on Quora, ask on forums, find out more and keep going.
• Do this for 6 months, non-stop every day. Then evaluate if all these machine learning stuff interests you.

If even after the 66 mini hell tour you are undeterred, you are now ready to face the 666 full hell formula.

1. From now on, spend 60 minutes every day on data science. Cover depth topics that interest you like NLP, recommendation systems, deep learning and complete MOOCs on big data technologies like Spark and Hadoop.
2. Do 6 data science-related projects a year. You can give talks, conduct data science 101 training for others, compete on Kaggle, build data products with open data, start a data science blog or start/join a data science meetup group, volunteer with data organizations, intern, work on data contract jobs.
3. Do (1) and (2) for 6 years. That was roughly how long I took, though honestly, I put in a lot more than 60 minutes per day so YMMV.

The 60 minutes you spend every day builds up your foundation. The 6 projects you do showcases your capabilities (be proud to share them during interviews). You will be surprised how much you know at the end of 6 years.

If you think about it, a Ph.D. takes 5-6 years to earn (more if you take Masters on top of it). So the 666 formula essentially trades those years for self-study and a wider range of experience.

Edit 1: Am I competitive with Ph.D. candidates? I have more hands-on experience and broader skill sets - from technical skills to understanding of business processes to recruitment. PhDs on my team have a more laser focus on their domain. I learn much from their depth, and they learn about my wide array of tools. I will say a good data team needs people with the breadth and the depth to function. I have huge respect for Ph.D. folks; I am just not the kind of person who can sit down and study one topic for 6 years.

Edit 2: One common follow-up question is: "If we need working experience to get hired for data science roles, how do we get the foot in?" Well, the good news is you don't have to wait 6 years to get hired. My path:

Intern at an analytics software vendor > junior engineer at big data hardware company > business analyst > data scientist.

With some foundation in maths and programming, you can get started right away, but it might take a while to earn a data scientist title - you might not be able to work at Facebook data science team right out of school, and that's perfectly normal.

Ultimately, the world is your oyster. I had met several masters friend who started as an intern at companies during their studies. Some joined startups for the challenge and exposure. You can also build up a portfolio via data contract jobs on portals like Experfy.com, competitions on Kaggle.com. Lots of smaller companies are looking for data talents these days - but you need to engage them in a conversation and to have a solid portfolio of projects help. See my answer to What's the best data science interview you've gone through?

Edit 3: I am now doing a computer science masters through Georgia Tech’s OMSCS program. Georgia Tech is famous for their artificial intelligence and robotics research, and I look forward to learning more under the Machine Learning specialization of the master program.

Edit 4: I am leading a small data team now, halfway done with my Georgia Tech Master and learning about deep learning. I still don’t understand many things I read about, but I am trying to get better every single day and enjoying the growth.

For more data related posts, visit my quora page.