Back

 Industry News Details

 
How One Nonprofit Is Using Data Science To Transform Lives And Communities Posted on : Mar 25 - 2015

Can data science save lives? Bayes Impact thinks so. The San Francisco-based nonprofit is using data for good to help governments and nonprofits tackle social problems.

A trio of entrepreneurs – Andrew Jiang, Eric Liu, and Paul Duan – left the private sector to found Bayes. Rather than helping technology companies convince more people to click on ads, the entrepreneurs decided to use their skills to create social change.

“If you look at the way organizations are run, the future is going to be more data-driven,” said Bayes Co-Founder and President Duan. “And we want to be the agent of that change.”

Founded in 2014, Bayes started out working with cities, including Seattle and San Francisco, to improve police responses to crime and optimize ambulance response times. They’re also addressing neglected areas of social change like criminal justice, education and health. From reducing prison overcrowding to improving academic outcomes, Bayes wants to change lives for the better.

The Power of One

Before starting Bayes, Duan often volunteered at homeless shelters and soup kitchens in his spare time but was frustrated with the limited impact he could have as an individual. “You can feel better about yourself for an evening, but it doesn’t scale,” Duan said.

The “aha” moment came when Duan joined Eventbrite as the company’s first data scientist, where he worked on algorithms to improve fraud detection. “That’s when I saw that by using data science you could scale dramatically,” he said. “If there’s one system that makes millions of decisions every day, and one person can write an algorithm to make that decision 10 or 20 times more accurate, you can affect millions of users.”

Duan and his co-founders decided to use the power of data science to make a similar impact in the social sector. The idea caught on: Bayes received a $50,000 grant from startup accelerator Y Combinator, and formed a roster of mentors and advisors from successful tech companies such as Airbnb, Intuit, LinkedIn and Netflix.

Thinking Big

To create the most impact, Bayes takes on a few large projects at a time rather than spreading its resources across many smaller projects. And rather than relying on volunteers, the founders decided to hire the best data scientists they could lure away from top tech companies.

Volunteer-led projects are great for raising awareness, said Duan. “But if you want to actually solve a problem you need a more professional way of going about it.  It’s not as easy as having someone do a hackathon and pretending you can solve Medicare,” he said.

The founders approached technology veterans who had already achieved career and financial success, and were interested in opportunities to make a difference. “We told them: Do you want to keep optimizing ad click-through rates for the rest of your life or do you want to come with us and change the world?” Duan said.

Lessons Learned

One of the nonprofit’s early lessons was discovering that there’s just as much red tape involved at the city level as there is at the federal level when working with government groups. The impact of federal projects, however, can be much greater, said Duan.

“I would say our best work is still ahead of us,” he said. “The real work we want to do is large-scale deployments with big nonprofits or at the federal level of government.”

Bayes also prioritizes projects that result in working operational systems to create more impact. “We try not to do projects that are purely research-based or projects that only result in a recommendation,” said Duan.

Scaling Social Change

Moving forward, the nonprofit is already working on larger initiatives, such as helping the Annie Casey Foundation safely reduce juvenile detention rates by a factor of two through its Juvenile Detention Alternatives Initiative (JDAI) that operates in nearly 300 counties nationwide.

Bayes is also helping the U.S. Department of Health’s Scientific Registry of Transplant Recipients make better matches between organ donations and people who need transplants. By better predicting the rate of organ rejections, the total supply of organs can be increased by as much as 20 percent, said Duan, saving thousands of lives each year.

 

“At the end of the day,” said Duan, “It’s really about making better decisions at scale.” Source