Database Reference
In-Depth Information
issues and in the regions where their organization works. And finally, we have
found that there can be communication or “translation” challenges when the
two sets of experts get together.
So a lot of the work that DataKind does is not only to recruit—bringing
the community together—but also to help make it as easy as possible for
people to work together and scope a project that will make a big difference.
We focus on ensuring that volunteer data scientists have the background and
framework they need to take an ethical approach to their work. We also work
closely with the beneficiary organizations, digging into the data to scope out
a project that will be useful and in support of their mission. Then all along the
way, DataKind stays close to the teams, as supporters, helping out when there
are roadblocks, providing camaraderie, and ultimately making sure the project
is a success for the volunteers and the beneficiary organization
We also see a large part of our role as the storytellers. We want to herald the
results and spread the word about the power of data science for good. We
hope that projects result in new tools or approaches for the beneficiary orga-
nizations, but also that the work done during the project can scale to benefit
other people and organizations.
For example, we recently worked with a nonprofit organization to build a tool
that looks at Google Maps to identify poverty levels in a Kenyan village based
on the buildings' roof types. In this case, we knew from the organization's
regional expertise that iron roofs connote more wealth than roofs that are
thatched. It's exciting to think that by using a computer-aided process, coupled
with knowledge about a regional culture, a nonprofit could save hundreds of
hours of investigation by foot, traveling village to village, and instead spend lim-
ited resources on programs to address pockets of extreme poverty. The work
was done for one organization, but the tool and approach has the potential to
benefit any number of organizations.
DataKind has been working to codify our process in a “playbook” so that, ide-
ally, anyone can replicate this type of collaborative work to harness the power
of data science.
Gutierrez:
Why develop a playbook for how DataKind does its work?
Porway:
From the very beginning of DataKind's project work, we've been
fielding interest from around the world. We have been fielding questions like:
How do I do DataKind in my city? How can I be involved in the Data for Good
efforts? I want to do it in my particular town—can you help me? We realized
from those requests that DataKind could achieve its mission of tackling the
world's toughest problems through data science by enabling others to do
data science volunteer work. They just needed the tools to ease their journey.
Now, three years into our operations as a stand-alone organization, we feel
like we have enough experience under our belt to share our framework and
process to support successful volunteer data science project work.
