Database Reference
In-Depth Information
just drawn in. What a crazy problem. How do you replicate something that
we can't understand ourselves? Leaving the philosophical answer aside, the
current thinking at the time was that machine learning—writing algorithms to
help computers learn models of the world from data—was the cutting edge
of approaching this problem.
I was accepted into the PhD program at UCLA in Song-Chun Zhu's lab to
continue working on the problem of modeling human vision. As a result, he
advised me to switch my major from computer science to statistics, since the
latter was going to be so much more instrumental to our work. That's when
things really changed for me.
During my time in the statistics department at UCLA, bolstered by internships
at Bell Labs and Google, I saw how my machine learning work could be applied
to different interesting problems. I saw that the world is changing under our
feet and that statistics will be needed everywhere. As one of my advisors,
Mark Hansen, pointed out, every field is going to need statistics and comput-
ing. I still believe that a data scientist is just a statistician who can program well.
Being able to collect data, model the data, visualize data, and draw meaning
from the data allows you to see in ways that you've never seen before. And
now because of the way data is coming off of everything, there are no limits
to where this thinking can be applied.
Gutierrez: What publications, web sites, blogs, conferences, or topics should
people read to learn more about the types of problems being tackled in non-
profit and social organizations?
Porway: First, I would say check out DataKind, of course. We publish all of
our case studies. We're also going to come out with our first failure report on
a project that didn't work, because we want to share learnings about every-
thing—not just the good, but also the bad. We're not shy about that. So defi-
nitely come and look at our case studies.
If you want to get your imagination going, the topic The Human Face of Big Data
has really good for examples of how the latest techniques are being applied to
interesting human and social problems. 1 There are a couple of other groups
that are definitely worth looking into. The group Markets for Good is looking
at how to design data markets to tackle social issues. Data and Society, Danah
Boyd's new group, is worth looking into because they are tackling questions
like the anthropological implications of big data. Also worth looking at is the
United Nations Global Pulse. They are dedicated to using data science to
tackle initiatives the UN identifies, such as coastal preservation informed by
satellite imagery.
1 Rick Smolan and Jennifer Erwitt, The Human Face of Big Data (Against All Odds
Productions, 2012).
 
Search WWH ::




Custom Search