a sociologist working at facebook

Michael Corey is a PhD candidate in sociology at the University of Chicago. This guest post explains his experiences working for Facebook, the world’s leading social networking website (as if you didn’t know that!).

Another Dispatch from Industry

Last summer I moved from Chicago to the bay area to work as a quantitative researcher at Facebook. I’d done six years in the PhD program at Chicago and left with drafts of all my dissertation papers but without a cohesive dissertation to turn in (3 paper dissertations aren’t exactly allowed). Six months at Facebook has been eye opening and weird. Below I’ll try to give readers a feel for what it is like to go from an academic track to an industry job.

The FB Culture:

The culture at Facebook is really fun. I work at the main campus in Menlo Park, where a few thousand people work on the various FB platforms and the associated companies (Parse, Onavo, Instagram, etc). My mother-in-law describes it as an Oxford College designed by Willy Wonka, which is pretty fair. The campus houses everything you need to reduce any external friction that would take you off-campus during the day [http://cnettv.cnet.com/barber-candy-shop-bank-among-deluxe-perks-facebook/9742-1_53-50153870.html]. It is pretty easy to drink the Kool-Aid about how great FB is, and I would imagine that it is hard to work here if you don’t. I wasn’t the biggest FB user when I started here, but having been off the site for a long time I learned to recognize how much I missed by not being on it. For so many of my peers it is the only medium to communicate news, baby pictures, or cat memes to weak ties. Risk taking is encouraged and speed is considered a virtue.

Communication:

Everything I’ve seen happen at FB is team-based. Teams of contributors work together and report up to a manager. That manager is in a team of managers, and upwards to the management team. There isn’t a lot of siloing, and you shouldn’t be tied to the whims of one single personality–a nice change from the pathological advisor-advisee relationships that can develop in graduate programs. Interdisciplinary work isn’t mentioned much here because it is assumed–everyday I talk with project managers, engineers, designers, researchers, and all kinds of other people involved in serving over a billion people each month. Everyone works together in big open spaces and being open about your work is a central tenant of the culture here. Practically, this means we can all get into each other’s code, results, and people constantly ask each other to consult on projects where applicable. A lot of this probably comes from the engineering influence. There are no closed doors and everyone will return your email.

The Stack:

My daily software stack is based in R, coded on sublime text 2, run on both my laptop (OS X) and my development server (CentOS). Within R I rely heavily on the ggplot2, plyr, and car packages. To produce pretty results in a standardized format I write markdown in knitR and produce JavaScript graphs with rCharts. Our data can be collected and managed via open-source tools built at Facebook, namely Hive and Peregrine. Both can be queried with SQL. One of my biggest challenges starting the job has been learning how to move from Stata-style merges to SQL, which is standard in almost every field but ours. Open source software rules at Facebook and most tech companies because it is free and customizable. Imagine having to buy Stata licenses for 500 data analysts! The only way that’d make sense is if you gained reputation by how much you spend–say how big your grant is! Big data is big, but there are amazing tools to work on it; once you get the data you end up analyzing a much smaller subset of cases. I know how mappers and reducers work, but I don’t think I’ll ever write one. Python is also widely used for data analysis and the marketers often use SPSS.

Academic Focus:

Facebookers are heavily involved with academic pursuits. We have a very active group of academics and coordination for this at the company level. There are deep links between HCI/Communication PhD programs and the data science team at Facebook. Data science has former professors on the team and hosts people on sabbaticals. They also keep a rotating stable of excellent interns. The majority of the team come out of HCI-type programs, but this is more a case of supply than demand, in my limited experience with the team. My own team (Growth Research) is made up of two sociologists and a manager trained in communications with a sociologist as an advisor. Many of the teams are doing social science though not always with the benefit and baggage of the formal social-science literature. Publishing is encouraged and various individuals produce work and referee for academic conferences and journals. We’re even hosting a conference on digital data collection the day before ASA; a call for paper should be released this week.

This is still a business:

Despite how fun it may sound above, this is still a business. I work everyday to expand the availability of the internet to developing countries. It is great and will hopefully reduce the digital divide. But it’ll also sell more advertising, like I see every day in my FB feed. I’ve never been involved with something that is in the newspaper every day; being in the SF bay area makes that more acute. SF is a messed up place right now. 30 years of intense NIMBYism means a housing shortage that is unbelievable and tech takes a lot of the blame for this, some of which is reasonable. Our company is in the S&P 100 and 500. I never expected to work for a multi-national corporation, let alone one with a (current) market cap double companies like Starbucks and Ford. While we are encouraged to publish, everything has to get approved by the communication and legal teams. It is a small price to pay for such an incredible data set; but we don’t do sharable academic data. This is less of an issue in sociology than in fields where open sourcing published data is standard.

Surprises, money, and misc:

A lot of stuff has been very weird for me. At 31 this was my first 9-5 job. I had been a freelancer between college and grad school, so I face a lot of surprises specifically because of that. For instance, Flexible Spending Accounts are amazing! I can budget money to healthcare *and* save taxes? Where was this when I was just getting by financially in grad school with terrible undergrad-level healthcare? Corporate discounts on rental cars and all kinds of random crap–who knew? I don’t have to work on evenings or weekends. Really. I pull a 40 hour week, depending on if I work during my commute. I’m given all the tools I need to do my job, including computers, books, space, time, and a ton of up-to-date colleagues. We get paid a fair market salary and even with living in SF it is a shock compared to being a grad student.

Conclusions:

There’s a ton of opportunity in tech for formally trained social scientists. Sociology is way behind the curve in terms of technology and that has been a key factor keeping sociologists out of the rising tsunami of data science. Handling massive streaming data is getting easier, as is teaching yourself basic computing. Companies are also reaching out more to social scientists to fill jobs that require a lot of inference, survey work, and qualitative work. There are lots of opportunities to learn on the job and for graduate students the internships at tech companies are an amazing and underutilized opportunity. I spend a lot of time recruiting, which is awesome and helps us get good people by recommendation. Yes we are hiring, send CVs to mrcorey@fb.com.

50+ chapters of grad skool advice goodness: From Black Power/Grad Skool Rulz