Frequently Asked Questions07 Dec 2013
I occasionally get emails from people asking me for advice or about my experiences before, during, and after graduate school. Here are some questions I’ve answered in the past. I’ll add more here as I receive and answer them.
Before Grad School
What was your experience like at the Fed Reserve Board and Matrix Group International, and how did they impact your decision to pursue a PhD in Information Systems at NYU?
I wanted to get a PhD in economics. At the Fed I did macroeconomic research and realized I liked research but didn’t like macro that much. At Matrix, I learned how to program professionally, which turned out to be a very valuable skill to have. I came to NYU because I wanted to study the social science of software development based on my experience at Matrix. I didn’t end up studying that, but IS is a very broad field where I eventually found a niche.
How did you choose to go to grad school? Or for that matter to go into the field of data science?
I wanted to be an economist in college, so my plan while working was always to go back to grad school for economics. It turns out I found a field that was a better way to merge my interests in economics, computer science and machine learning, so I ended up in Information Systems (which is a field offered by some business schools). For me the decision was mostly about wanting to know some subject incredibly well. It’s possible to do that on your own, but the structure/community that grad school provides makes you way likelier to succeed at becoming an expert. I didn’t choose “data science” but what I ended up learning was a really close fit to that nascent field.
If I had to start over, I’d get a PhD in Statistics instead of Information Systems. Though I love the social science research that I get to do, I think statisticians have the most impressive toolkits for approaching problems, and are definitely the most employable :)
About NYU Stern and Applying
What attracted you to Stern?
The professors in my department are leaders in our field. After meeting with them and hearing about their research, I knew that coming to Stern would give me the opportunity to work on something groundbreaking.
And now that you’re at Stern, what are some of the things you like best about being here?
Stern has a vibrant research community. Just about every week, there’s a student or professor giving a fascinating seminar. The faculty are eager to meet with students to discuss ideas and help out with research. Because we’re in New York City, we are able to attract some amazing visiting speakers from all over the world. Within the doctoral student community, I’ve made some great friends who are incredibly helpful and supportive. It’s nice to share your challenges with people your respect and admire.
How have you grown as a student/researcher since being at Stern? What allowed that to happen?
The doctoral program provides a steady stream of challenges, from introductory classes to qualifiers and dissertation work. It shapes the way you think, changing what you think is interesting and how you approach problems. There’s no single event I could point to, it’s a cumulative change. You learn that almost everything is more complex than you believe at first. When you accept that, you can research with an open mind and really make a contribution.
How would you describe the people or community (whether it be NYU, Stern, the doctoral program, or your dept)?
The professors in my department are friendly and supportive. They understand what you’re going through and are always forthcoming with advice. I’m obviously closest with my cohort in my department–I’ve shared an office and taken most of my classes with them for three years now. We’ve built a strong bond by going through the same challenges again and again. It’s a real privilege to learn and work with such talented people.
Does the stipend include a certain amount for rent, or does Stern have specific housing for PhDs? How has living in NY been like on the stipend, have you needed to supplement it?
No the stipend is just a payment you receive twice a month. You can spend it on housing anyway you want. I think for a limited number of students there is NYU housing available, but it is not super common for US students to receive it.
It is a bit difficult to live on the money, but you actually work very hard so there isn’t much time to spend. After you pass your qualifier you can take a small loan each year, so between that and the stipend, I have been fine. It is not possible to work during the program except during the summers as an intern.
Is there anything else you think it would help me to know about before applying to grad school?
Enjoy not being a graduate student before you start. It is a major life change. It’s not bad, but you’ll be busier than ever and your hobbies/interests/relationships will all have to evolve as a result.
Should i mention names of specific profs in NYU Stern in my SOP? If yes then whom should i mention about?
I wouldn’t mention any specific professors unless you can write cogently about some of their work which interests you. I would encourage you to read their websites and see if any of their research is something you’d consider building on.
How should I substantiate why I want to study in NYU Stern?
Make sure you know what our department does best (look at faculty websites) and make the case that you have similar interests to our faculty.
Any other admission tips would that you consider can help a prospective International Applicant?
Make sure your personal statement is thoughtful and shows some broad research interests. The faculty look for technical skills and also passion for a specific topic or area. You don’t have to know exactly what you’d like to study, but I would write about what questions you’d like to answer as a researcher.
What specifically does the admission committee look for in successful applicants for the Information Systems stream?
What are the first two years of grad school like?What are the first two years of grad school likeasdff?I haven’t been in the meetings where they admit students. However, the three keys seem to be 1) technical proficiency, 2) intellectual curiosity and 3) a good fit between your interests and the department’s.
What are the first two years of grad school like?
For the first two years, you generally spend the bulk of your time on coursework, much like a masters. Some courses are very hard and require a lot of time. Others are lighter and just meant to provide some familiarity with some topics, revolving around class discussion (surveys or seminars are like this). I’d say you can expect something like 10 hours in class and 40+ hours of work outside of class, mostly reading and then writing papers. This can be frustrating, but it’s pretty vital to be “caught up to the current conversation”–having the background knowledge required to have a cogent discussion of the interesting questions and issues in your field. A HUGE part of academia is just knowing how things are defined, what people already know, and what things they think are interesting. If you’re like most people, you’ll come out of this period realizing that what you previously thought was interesting has already been studied pretty thoroughly, so you’ll have settled on some new questions you’d like to address in your research. It’s possible and normal to end up somewhere pretty far from what you expected.
What kind of demands are there on your time?
In my program I had to work on research alongside my course work right away. I would characterize this as exactly the juggling act you described. Faculty members put pressure on you to finish projects with them. You have reading and assignments for course work (which can continue into your 3-5th years as you address subjects you feel weak in). You tend to have some papers or projects for courses or program requirements that also are very time consuming. You’re expected to attend 1-2 seminars a week with visiting speakers or members of your department. In addition, you’ll always have some papers and books you feel like reading because they fit your interests. Some people can really find some synergy between all these requirements, for instance writing their course papers about things they eventually want to research, so as not to duplicate reading. I’ve found this to be really difficult. The point is that there are many demands on your time and it is usually overwhelming. It certainly detracts from your ability to focus on ideas you’d like to work on.
That said, you would be hard pressed to find a career path that encourages independent and creative thinking as much as a PhD. The reward for being stretched in so many directions it that the important deliverables are always your own ideas, though they must be written and researched very carefully. I’d say most people have a good idea of the benefits of a career in academia, but they don’t realize a) how much of your time it takes just to be a functioning member of a doctoral program (the stuff I wrote about above) and b) how time-consuming and difficult producing quality research is. Really a doctoral program takes up way more of your time than you’ll ever imagine beforehand. Unlike in a job, you’re rarely completely done with anything and it’s always possible and desirable to put more time in. One of the hardest things is knowing when to stop and go back to normal life, or even feeling like you’re allowed to do so. Everyone I talk to about their programs admits to me that they have had to stop reading for fun, watching much tv, hanging out with friends, and pursuing the hobbies they had before.
How do you know if a PhD is a good fit?
I genuinely enjoy the challenge of the process and others do, too. Some people in my program seem to have made a mistake. I think the key difference is that for them, the work seems like an obligation instead of something they volunteered for. It’s easy to get caught up in meeting requirements and what it takes to finish, but really you are signing up for a lifetime of this stuff. If you don’t like the day-to-day stuff, which requires juggling a few tasks simultaneously, it’s will be tough to be happy in a program like mine. In fact, I think they go to great lengths to make you do a lot of things at once to get used to what it’s like being a professor.
How long before you should start to publish articles?
Publishing is different by field, so I’m not sure how much advice I can offer there. I’m entering my 5th year and I only have conference (no journal) publications, but that’s not very unusual. My other general advice is all about thinking long term. There’s never as much pressure as you think there is, and it’s something to be enjoyed (or else why would you spend 5+ years doing it?). The real key is finding a balance where you can be happy and productive, while fulfilling what they expect of you to graduate. I think many people miss this point and see the goal as “getting the most papers” or “getting the best job,” when really what you are optimizing is your own life. Early on, the key is generating an idea of what you love to work on and find interesting enough to devote time to later. The happiest/best researchers are those who have some problem they love and are dying to tackle. They find this problem early in grad school (or even enter with the idea). So basically find out what’s interesting to you early on (and prize this above getting good grades, which barely matter). The least happy people are the ones who are trying to think of a research topic (or just pick one arbitrarily) so they can start/finish their dissertations. There’s also a host of advice to be given about choosing an advisor, but the gist is pick someone you personally like, does what you want to do, and isn’t overloaded.
What skills, both technical and otherwise, do you find most valuable for the research that you do?
Asking good questions is the most important thing. This is a skill you develop by talking to people in a field of research for a long time and realizing what are the big open things that people don’t understand very well. Think about the problem a journalist faces. You want people to read and talk about what you write. If you can’t make people interested, all the analysis and data collection will go to waste anyway.
Besides that, programming is incredibly important because you always have to manipulate data in some way. Statistics is also a huge part of what I do. You can never know too much of either of these things. I’ve recently gotten into data visualization, which is an underrated skill.
It seems to me that the largest obstacles to new data creation/collection are large fixed costs for an investment (research) with an unknown return. As a result, you typically need to have external funding. How do you think an individual can best overcome these obstacles for conducting an individual research project without external funding?
It’s a bit of a cop-out, but I would say creativity is the important part here. The best research I read, I often think “wow that was brilliant to get data in that way.” It’s not so much the analysis, it’s the idea to gather data from somewhere. As an example, when I was at the Fed in 2004, I had the idea to scrape Craigslist to get apartment prices in various cities. I never ended up working with that data, but I know that could have turned into a big research project if I had followed through.
One specific tip I will give is that often the best way to create novel data is to combine two sources people hadn’t thought of joining before.
As someone very interested in applied behavioral research (for a number of different topics ranging from the NFL to public policy), are you aware of any opportunities to become more involved in this type of work?
There are plenty of jobs out there if you are looking for formal employment. Your best shot is to work for a small company or a startup that will give you the leeway to learn on the job. Bigger places are looking for people with prior experience, so you’ll face a cold-start problem there. Plus they may not give you much room to shape your own job.
If you want to pursue it as a hobby, there are options like volunteering for a project at DataKind, or finding a hackathon which is data-oriented (I would look on meetup.com). You really need to be plugged into a community of researchers to do this kind of thing, because it’s hard to do any projects in isolation. In NYC, this revolves around several meetups which many of the data scientists in the area attend. There’s also academic communities out there and they are usually happy to let interested people show up to talks. Ask to get on mailing lists for visiting speakers if you are near a university and know a student there.
How did you get involved with Facebook, and what motivated you to decide to do your research on NFL/likes?
I interned at Facebook last summer. I proposed some of my dissertation research to them, applied for an internship, and got the job. I did the NFL work because I like the NFL and I wanted to write a blog post to get some of my work in front of people. Academic research doesn’t have as broad a reach as small data projects like my NFL analysis.
What is like doing research at Facebook?
Facebook is a great place to work or do research. They are very supportive of asking and answering hard/interesting questions. You can observe a lot of what a large number of people are doing online. The downside is that you must study human behavior on the web, which kind of limits the questions you can ask (e.g. you can’t say much about consumer behavior past clicking on ads because you never observe purchases).
The Data Science team at Facebook does top notch research on par with most academic departments that study similar areas. The size of the data isn’t really the major advantage, it’s the richness. It’s very fine grained, so you can answer many questions that were not possible before.
What was a day in the life like as a data science intern?
It’s a pretty sweet job. You work on problems/projects that are anywhere from 3-months to a year. They tend to be deep dives into problems/questions/techniques where the company would be interested in the results. They are higher risk than most projects because you’re trying to go for a big impact and novel work. There are some shorter-term projects mixed in that involve answering more immediate questions, and you tend to consult with other teams on problems they have where they don’t have enough expertise. Basically, you’re paid to think really hard along with a bunch of other smart people and collaboratively come up with completely new stuff. So that’s obviously pretty great. You also get a lot of choice in what you’d like to work on because there are usually more than enough problems/projects to choose from.
What is the career of a data scientist like? What happens after that first entry level job? This is pretty new ground, I think. I honestly don’t know what the field will look like in a few years. I see demand increasing, so there will be more entry level work or tasks that start to be done by MBAs or software engineers, but the more advanced stuff is not going away either. I would view the career as a series of new challenges. You want to work somewhere that gives you *hard* problems because those are the interesting ones and also where you can differentiate yourself. You get good at one kind of problem and it’s not hard anymore (or you solve it with technology), so you move onto new ones. I have friends that switch jobs just so they can work on something different. You could picture kind of hopping around from place to place and accumulating experience in different domains and trying to make the biggest impact – that’s how I see my future career path.