Databases Reference
In-Depth Information
University. David spoke about Google's approach to social research to
encourage the class to think in ways that connect the qualitative to the
quantitative, and the small-scale to the large-scale.
Google does a good job of putting people together. They blur the lines
between research and development. They even wrote about it in this
July 2012 position paper:
Google's Hybrid Approach to Research
.
Their researchers are embedded on product teams. The work is iter‐
ative, and the engineers on the team strive to have near-production
code from day 1 of a project. They leverage engineering infrastructure
to deploy experiments to their mass user base and to rapidly deploy a
prototype at scale. Considering the scale of Google's user base, redesign
as they scale up is not a viable option. They instead do experiments
with smaller groups of users.
Moving from Descriptive to Predictive
David suggested that as data scientists, we consider how to move into
an experimental design so as to move to a causal claim between vari‐
ables rather than a descriptive relationship. In other words, our goal
is to move from the descriptive to the predictive.
As an example, he talked about the genesis of the “circle of friends”
feature of Google+. Google knows people want to selectively share;
users might send pictures to their family, whereas they'd probably be
more likely to send inside jokes to their friends. Google came up with
the idea of circles, but it wasn't clear if people would use them. How
can Google answer the question of whether people will use circles to
organize their social network? It's important to know what motivates
users when they decide to share.
Google took a
mixed-method approach
, which means they used mul‐
tiple methods to triangulate on findings and insights. Some of their
methods were small and qualitative, some of them larger and quanti‐
tative.
Given a random sample of 100,000 users, they set out to determine the
popular names and categories of names given to circles. They identi‐
fied 168 active users who filled out surveys and they had longer in‐
terviews with 12. The depth of these interviews was weighed against
selection bias inherent in finding people that are willing to be inter‐
viewed.
