Daniel Tunkelang - Data Scientists at Work - page 91

Gutierrez: How do you decide what to read and what startups to look at?
Tunkelang: These days I rely heavily on my professional network to curate
what's out there. I check my LinkedIn feed, naturally. I also see what the people
I follow share onTwitter. And I use a few news aggregators—mostlyTechmeme
and Prismatic. As a result, I get a good mix of front-page technology news and
niche information, and I do my best to keep up.
Gutierrez: What types of data are these data scientists working with?
Tunkelang: All kinds—text, numbers, clicks, relationship graphs, geography,
time series, and similar data. Part of the challenge of working in data science is
that you tend of have lots of different kinds of data, and it's your job to stitch
them together.
Gutierrez: Are there data that data scientists are not yet looking at?
Tunkelang: Probably not. But I think there's a lot of work to do on improv-
ing how wearable devices collect data in order to truly deliver better living
through data. I'd love to have a device that tracked everything from my physical
activity to my mood, which then allowed me to benefit from data analysis
without risking the exposure of private data that is deeply personal. We're
getting there, but these are early days for wearables and sensors in general.
Gutierrez: What resources are helpful to your work?
Tunkelang: Since I focus on search, I keep up with the leading information
retrieval conferences, in particular the SIGIR and CIKM conferences spon-
sored by the ACM. I also keep a foot in the big data world through con-
ferences like O'Reilly Strata. I don't read particular blogs anymore. Rather, I
mostly rely on LinkedIn and Twitter to surface relevant content to me. I tend
to use topics mostly as references. Fortunately, some of the most valuable
information is timeless.
Gutierrez: How do you keep track of things when you are out of the office,
given the ephemeral nature of feeds?
Tunkelang: Sadly, I'm almost never offline, as I'm a bit of an information
addict. I did disconnect for a week recently and I spent most of the following
weekend catching up. For all of the effort that has gone into news aggregation,
there's certainly more emphasis on providing real-time feeds than in aggregat-
ing feeds over several days, let alone several weeks.
Gutierrez: What does a typical day at work look like?
Tunkelang: I'm not sure there is a typical day. Broadly, I lead a great team and
then spend the rest of my time on hiring and outreach. Most of my time is
spent providing guidance to my very accomplished team—adding value where
I can, and staying out of their way where I can't. I also spend a lot of time on
hiring and outreach.