#TechTuesday: How Non-Profits Can Save Lives with Code

shutterstock_60670483.jpg

As rescue efforts continue in Nepal, discover how lives have been saved in previous crises through code. By Erin Boehmer (Master of Data Science Student, UC Berkeley)

If you Google “Haiti Earthquake,” you’ll find search results scattered with devastating pictures of crumbling homes and chaos. What you won’t see is the beginning of a crisis response revolution.

In 2010, technical volunteers gathered around the globe to mine Twitter for tweets that would be useful to aid workers on the ground. They developed and deployed apps to help find missing persons.

They worked through the night to map everything from water shortages to vandalism from translated Creole text messages. In a chaotic environment where communication could be the difference between life and death, hackers came to the rescue and saved lives with code.

Saving Lives Through Data

As an undergrad with the skills of a web and mobile developer and the heart of a humanitarian, I was instantly hooked on this effort and spent a summer studying technical crisis response volunteer networks at the Woodrow Wilson International Center for Scholars in Washington D.C.

The nonprofit sector is rife with projects that operate as though the developing world is Kevin Costner’s cornfield from Field of Dreams: if you build it, they will come!

Well, long story short, that’s not how the world works. But when the Haiti earthquake shattered people's lives, it also clarified their needs and demanded humanitarians become responsive rather than prescriptive.

How did the techies then step in to create smarter response systems? By gathering, analyzing and creating action from data.

I love data. Data speaks across cultural divides and gives a voice to the individual. And, as a Master of Information and Data Science student at UC Berkeley, I’m starting to listen.

Data – especially “Big Data” – is already all the rage in private companies from Silicon Valley to Wall Street. But I don’t hear as much about data-driven decision making in the nonprofit world.

Nonprofits and companies in emerging markets can use big data, in combination with machine learning, to reveal exactly what people need so that aid efforts and new products are effective. In order to realize that power, however, you need to watch out for some common pitfalls.

Pain-in-the-Butt Data Collection

A lot of development projects expect buy-in from the people they help because it’s “for their own good.” It’s just filling out a survey or form to get feedback from users so that medical attention, education, clean water, etc. etc. can be provided more effectively.

But who likes filling out forms? Theoretically healthcare forms are for my own good, but if I’m not dying, I avoid that tedium at all costs.

The same goes for development and aid recipients. During the Haiti earthquake, people were desperate and willing to spend extra time channeling their attention towards creating data for crisis responders.

In a non-urgent situation, don’t expect this to be the case. Surveys and forms are slow and hard to collect and largely unreliable if you’re depending on governments in developing regions to do the dirty work.

At times, these slow methods may be necessary, but ideally you want to find or create a reliable source of data exhaust produced through transactions and services people are already using. Make your data collection systems smooth and invisible whenever possible – so long as you’re not violating privacy rights, your users will thank you.

What is Your Data REALLY Saying?

I love data and believe in its power, but also know it has limitations. As economist Ronald Coase phrased it, “If you torture your data long enough, they will confess to anything.”

But it's not just mistaken mathematical assumptions and convoluted models that can confound the truth; it’s also assuming that your data represents the truth. We may need data about what people are thinking, seeing and doing, but we can’t markup a human to transfer through some protocol (yet!)

So we make do with what we have: tweets, blogs and a lot of fancy models. In the end, though, they’re all models that are dependent on the limited data we give them. So before you decide to do another Twitter sentiment analysis on the meaning of life, just think about the reality of what thoughts and ideas people actually express in 140 characters and adjust your expectations accordingly.

Is There an Ecosystem?

It’s really hard to grow an idea successfully if there is no ecosystem. When I was a first year in college, I worked on a water filtration project in South Africa to try to provide clean water to two rural villages. My team considered all the latest development buzzwords: we had “community partners” and a “sustainability plan” to maintain our slow-sand water filtration system! Fool-proof!

Except, as it turns out, the village chief thought the water committee we had created threatened his authority. So he disbanded the group. Within a year, the filter was a glorified puddle.

What I learned from both my experience in South Africa and with Haiti: sustainability isn’t something you write up in a Word Doc at 2 a.m. Lasting change has to come from within a community.

When people recognize the need or want something and are willing to change their behavior to get it, that’s when you’ll be able to start gathering the data exhaust that makes this ecosystem thrive.

How else have you seen data translated into action?

Photo credit: arindambanerjee via Shutterstock.


About the guest blogger: Erin Boehmer is a student in the UC Berkeley School of Information obtaining her Master of Information and Data Science. While currently serving as a SMART Scholar through the Department of Defense, Erin has contributed to both local and global social impact projects through groups such as Code for America and the United Nations Office for the Coordination of Humanitarian Affairs. Erin graduated as a Rodman Scholar from the University of Virginia in 2013 with B.S. degrees in Systems Engineering and Computer Science.