Database Reference
In-Depth Information
cheating somehow because there often wasn't a beautiful model of “reality”
behind the calculations. But now I've come around to think that simple models
are so much more useful, given all the volume data now available. I used to say,
“If the model doesn't explain
how
the process works, you didn't work hard
enough.” Now I'm like, “Eh, it's pretty functional. Just build a discriminant func-
tion over some data. That's good enough.”
I want to highlight another thing that I've changed my mind about. As much as I
wanted to use my skills for good, I actually did think that data and data science
were mostly going to be applicable to tech problems and tech solutions. What
has really shocked and surprised me in a good way is that there's almost no
limit to where data and data science can be applied.
Gutierrez:
How do you measure success for yourself and for your
organization?
Porway:
I have tried to align DataKind with my own values. I want to cre-
ate something that makes a significant improvement on the world. That's the
ultimate success for me.
At DataKind, we're challenged in a very practical way in figuring out how we
should measure the success of our projects. We measure everything from
how much improvement we've seen in organizational efficiency because of
analytics and data science, to how much more data literate an organization
is. Ultimately, we better at least leave this place a little better than we found
it, and that's only going to happen if we help each individual organization we
work with get better at what they do.
Gutierrez:
What does the future of data science look like?
Porway:
The data moment is now—and data literacy in the public is going to
be a requirement going forward. Right now, data literacy is like literacy in the
old days. Only the monks used to be able to read and write. That's dangerous
because it means the monks, even with their best intentions, are acting from
just their own worldview in terms of what gets written and what gets read. It
was only once everyone was taught to read that we had the beautiful range of
communications where people could share ideas.
I think right now, similarly, data science is in the belfries of academia, big com-
panies, and Wall Street. We need to smash those silos to the point that every-
one else can at least converse about data science. If we don't, we as a citizenry
risk losing the ability to critically assess data-driven results, which means
we can be manipulated, taken advantage of, and cut out of the conversation.
As governments and companies move to use data science more heavily, data
literacy almost starts to feel like a civil right.
Gutierrez:
How would you compare where we are now to four years ago?
And where do you think we'll be four years from now?
