The Data Incubator is a course for PhDs who want to exit academia and enter into the world of Data Science. We talk to Jon Ferris, a cofounder and Director of Partnerships at The Data Incubator, about the unique 6-week course, requirements for admittance, and how the "bootcamp" model is being applied to Data Science.
What is your role at Data Incubator? How did you get involved in the bootcamp space?
I’ve developed a passion for big data over the past couple years, and I was brought on a few of months ago to join Michael and help create the Data Incubator. Michael is an academic, a data scientist, and also one of our instructors. He’s focused on building the curriculum and speaking at conferences. My speciality is more around developing partnerships and growing businesses.
Where is Data Incubator teaching classes?
We’re working and teaching out of the AlleyNYC campus in Manhattan.
This will be your second course?
Yes, and it starts on September 9th
Are you all getting a lot of applications?
We had about 1,000 for our summer fellowship, probably because it’s free for fellows. Appliants need a PhD to apply, along with coding skills, strong aptitude for math and statistics, and the ability to communicate effectively.
What kind of applicants are you looking for? The website talks about applicants needing to have 90% of data science skills already. How do you know who qualifies?
They have to have a strong working knowledge of mathematics and statistics. They have to be able to code in at least one object-oriented language, and they have to be able to prove it; the admissions process includes a series of challenge questions, a code review, and an oral interview.
What’s the structure of the boot camp once students are there? What do you fit into those 6 weeks?
It’s an intensive so it’s onsite, in person, in New York. The first part of the day is spent in lecture, going over best practices for a particular data science tool or technique.
At the beginning of the 6 weeks, Fellows are asked to create a portfolio project. This is where they decide how they want to showcase their talents to employers. The project involves manipulating a large data set to come to some sort of conclusion. How do crop yields in Pakistan affect equity trading in Manhattan? So we have lecture in the morning, lunch, portfolio project development in the afternoon, and then at the end of the day, every day we have an employer come in.
Tell us about those employers.
During the program we have employers come visit the Fellows every day. For example, someone from Microsoft will explain what they’re doing, some challenges they work on, what data they’re working with. And they’ll answer all of the fellows’ questions. The fellows get great exposure to hiring companies and of course it gives the employers a sense of the Fellows, before they ever decide if they want to interview them. It kind of breaks the ice.
Why are you confident that the fellowship model can be applied to data science as a subject?
These folks are in really, really high demand right now. It’s very competitive to get them to come work for you. But you can’t create a data scientist overnight, and certainly not in a 6-week program. So everybody we’re working with has 5 to 6 years of advanced academic training. The chance of becoming an effective data scientist is much higher.
It doesn’t mean that there aren’t self-taught Data Scientists out there – there are. There are folks who probably don’t even have college degrees who work as data scientists because they were too smart or didn’t have the patience for an advanced degree. There are some terrific data science programs popping up around the country - just be wary of boot camps that say they can create a data scientist overnight.
We’ve started to see schools like Metis launch data science programs which are 12 weeks and you need to have some prior experience in Python, but you don’t have to have a PhD. What do you think about schools like that? I think the field is so broad that there’s potential at any level. You could take an entry level business person and put them through a 12-week analytics bootcamp and they can speak knowledgeably about that subject.
So I think it would totally help them in their careers and help them get a job. Any education is better than no education. There’s a real variety out there. Berkeley launched their Master’s program last year- it’s $53,000. Then you have some bootcamps for data science that are 12 weeks that are about $15,000. And then there are $6,000 programs on Coursera. So there’s all kinds of options. This is a great time to pursue education in this space.