Human-centered data

1 minute read


I just read two articles from my DataScienceWeekly email (so good! You should subscribe!) which do a really good job of humanizing data, talking so respectfully about its potential downfalls while also recognizing its tremendous opportunities for impact.

Deep Fried Data

Of course, I love the call-outs of bias and “dirty oil” in machine learning, but my favorite part of this speech was the deep respect for humanity’s messiness:

What irks me about the love affair with algorithms is that they remove a lot of the potential for surprise and serendipity that you get by working with people.

What if somebody grabs all this data, and does something with it that’s not scholarly? Well, that’s what you want! A sign of life!

Moving from real-time data to real-time programs

This is a blog post from Reboot, a consulting-ish firm focused on working with governments, foundations, and non-profits. The Aid Work panel at the Data for Good conference I went to a few weeks ago was moderated by Reboot. I like this blog post’s focus on “putting in the work” - collecting data isn’t always the answer, and even when it is part of the answer it’s not usually the hardest part. I’m also so happy when I see organizations approach conversations as inherently valuable - it was like when I discovered the Public Conversations Project during conflict training, and I realized that it was totally valid to think that conversations have inherent value and potential for deep impact.

“The development programs that are most precisely and easily measured are the least transformational, and those programs that are most transformational are the least measurable.”

So when we get excited about new technology and new data, we also have to get excited about the processes, the time, and the conversations we’re going to put into extracting value from the data.