At a data science conference in Berlin last year, a panel discussion inevitably turned towards the hot topic of data science for social good. Many of the panelists offered the same platitudes that I’m sure countless of you have heard before; it’s an emerging field. Great innovation will eventually happen. The future looks bright for data science and philanthropy. But one retort (by Klaas Bollhoefer of the unbelievable Machine Company) stood out. It stayed with me. It’s been the basis of many subsequent discussions with data science professionals. His response was two simple words:

“Who cares?”

Bollhoefer was not dismissing the idea, or suggesting he himself doesn’t care. He was asking, “Who cares?” Who are these people? Who’s actually going to dedicate the time to changing the world with data science? It’s an uncomfortable truth, but a truth nonetheless. When we think of data science, we don’t think of charities, NGOs or civic organisations. We think of Facebook. We think of Google. We think of Palantir.

This isn’t a mental association that looks to be erased any time soon. For young graduates with data science skillsets, success is a desk in the Google campus. It’s a healthy paycheck from a world-renowned company, where their work has an impact on our everyday lives; even if that impact is making us click an ad. This is absolutely not to paint data scientists as somehow more selfish than the rest of us- the vast majority of people in any profession can’t and don’t put charity first. Many data scientists do take on side-projects, but the part-time nature of such initiatives stunts their progress. These unfortunate truths mean the people with the brains and the skills to change the world are dedicating most of their time to less meaningful and more profitable projects.

But the story isn’t all doom and gloom. There are organisations trying to change this. One such organisation is Bayes Impact, a non-profit group based out of San Francisco, aiming to bring together top talent from the world of data science to work on social problems. They made a big splash last year, launching with a $50,000 grant from Y Combinator and a burning desire to save the world. Now, close to celebrating their first birthday, they’ve put their name to an impressive roster of projects, including fuelling Parkison’s research, optimising San Francisco’s Fire Department and detecting fraud in microfinance.

As recent fellow Stephan Gabler told us, Bayes Impact want to “show people that data science can be used for more than selling ads.”

“The fellowship is based on the assumption that the biggest impact can be achieved by having people work full time on a project,” he explained. “We usually work in teams of two fellows on a diverse range of independent projects.”

“Bayes provides a very creative and inspiring atmosphere in which the fellows work on their individual projects, that they also manage by themselves. Everybody is experienced enough to work completely independently, but we also all share our knowledge and learn a lot from each other.”

Gabler wasn’t following a “path that had data science as a goal from the beginning”. Like many who go on to be prominent data scientists, it was academia, research and problem solving that first appealed to him, in particular the area of “quantitative analysis of complex systems”. His path followed the typical progression of many data scientists; academia, more academia, a prestigious job afterwards. “I did an undergraduate degree in cognitive science with a focus on computer sciences, neuroinformatics and AI. During a year abroad at the Hebrew University in Jerusalem I came in contact with computational neuroscience.

“After finishing my first degree I went to join the Bernstein Center of Computational Neuroscience in Berlin and graduated with a masters degree about two years ago. Directly afterwards I joined patience.io as a data scientist. Patience is an adaptive online learning platform that uses machine learning to personalize the learning experience of its users”- a job that harnessed his years of experience with cognitive and computer sciences. As a resident of Berlin, I can confirm it’s one of the cheapest western capitals. As a data scientist here, we can presume Gabler was living like a King.

But of course, this is the point in tale where Gabler’s story deviates from the norm. At the beginning of the year, he moved from Berlin to San Francisco to join the Bayes team. I was keen to understand why Gabler bucked the trend; to paraphrase Bollhoefer, why did he decide to care?

“I always wanted a job in which I could apply my skills to something meaningful,” he told me. “Thats why I joined patience.io after graduation because I believed that online learning can affect many people in a positive way. At Bayes Impact I have even more leverage to have direct impact on social problems.

“Additionally, San Francisco is not the worst city to live in.”

It may not be the worst, but it is one of more expensive. Thinkpiece after thinkpiece has been penned about the skyrocketing prices and gentrification around Silicon Valley and beyond. Gabler admits “The fellowship is not comparable to a salary at Google or Facebook, but it is enough to live on.” Of course, all of this costs money- Bayes Impact are currently holding a “philanthropic seed round”, asking the public and investors alike to donate to keep the brightest minds in the business working for the greater good.

I ask Gabler if he considers the field of data science for social good stunted by the fact so much of the top talent is steered away from passion projects by the tech titans and their huge salaries. “I see this problem, but thats exactly where Bayes Impact comes in,” he responds. “We provide people who are interested in working on social problems a platform where they can do this in an effective way.”

Of course, for those with a data science skill set who want to do some good, Bayes provide the perfect platform. Gabler is currently working on an organisation called Youth Villages. “They are a big non-profit organization with the goal to provide help to troubled kids, focusing on intra-family treatment”, he elaborated. “They are an amazing organization with over 25 years of experience.”

The link between this organisation and data science may not be immediately obvious, but rest assured Gabler & co. are definitely putting their expertise to good use. ” They sit on a treasure of medical records and we can help them making their treatment even more efficient by analyzing this data with the goal to of building predictive models based on the historical data. This can help us to personalize the treatment of these kids.”

Gabler’s work has a direct impact on the lives of troubled and disadvantaged children. If we can convince more of the best and brightest that this alone is more rewarding that any paycheck, data science for social good may have a future yet.

Image credit: Bayes Impact

At a hackathon hosted by Bayes Impact, data scientists pitted against each other to develop data driven and implementable software solutions for high impact social problems. The winning hack will become a Bayes Impact project, staffed with full-time data scientists and engineers who will work with their partners to bring the hack to life.

Bayes Impact is a Y combinator- backed nonprofit that deploys data scientists to solve big social problems with civic and nonprofit organizations. They run a 12-month full-time fellowship for leading data scientists and work with civic and nonprofit organizations such as the Gates Foundation, Johns Hopkins and the White House.

Co-hosted by Jerry Yang’s AME Cloud Ventures and Joe Lonsdale’s Formation; more than 100 data scientists, engineers and designers at the OpenDNS headquarters were given 24 hours and access to a large amount of data to create an app for a social cause.

The winning hack was a web app that detects prostitution rings by monitoring where adult ads are posted. The team used data from Thorn, a nonprofit that aims to fight child sex trafficking with technology. In second place was a web app that used medical data from the Food and Drug Administration to predict interactions between prescription drugs, even if the combinations have never actually been tested.

A team of Facebook employees, who called their project Out For Justice, developed an app that used crime data to better prioritize 911 calls and place patrol cars for the San Francisco Police Department. Their project won third prize. (source: mashable)

Challenge-specific winners for the event were- Donors Choose project: Team Insight; High Point Police Department project: Team Out for Justice; White House Department of Labor project: Team Mine Risk Evaluator

Facebook behavioral economist Alex Peysakhovich said ‘s very impressive that people took this huge data set and did something crazy with it and said ‘here’s something useful you might want to know.'”

Jeff Chung, a principal at AME Cloud Ventures who was on the judges panel, said he thought data projects come out of the competition more fully realized than those at a conventional hackathon.  Chung said the projects showed a surprising amount of range, given that many were based around the same sets of data.

“The data hackathon may seem like it may restrict the products that come out of it because you are setting a foundation, but it’s really interesting to see how people see things differently,” Chung said.

Andrew Jiang, Paul Duan and Eric Liu, the three founders of Bayes Impact, are entrepreneurs come from diverse backgrounds. The engineer, data scientist, and investor had previously worked on or analyzed investments in some of the toughest problems facing all technology companies in the Internet era — how to get more people to click on more stuff.

But in recent years they’d turned their attention away from advertising click-through and to other issues, like criminal justice reform, fraud detection among micro-lending platforms, and better and cheaper research into some of the potential causes of Parkinson’s disease. Andrew said the founders first discussed the idea behind the nonprofit at another hackathon-for-social-good hosted by Paypal in April. Jiang said the founders saw a potential for data scientists to use their skills to improve nonprofit efforts that wasn’t being realized.

It was their belief that the data science which allowed them to move more units could be focused on pressing problems facing civil society, and as with click-through rates, someone would hit on a solution.

“We don’t do this as a service for this or that organization,” says Duan. “We focus on one area and come together to say how can we make this issue better as a whole.”

Read more here.

Follow @DataconomyMedia

(Image credit: Bayes Hackathon 2014)

Imagine knowing which buildings would catch on fire, which people wouldn’t pay their loans, or to solve overcrowding in California’s prisons. Only a few years ago, such ideas might have seemed outside the realm of possibility. Now, the Bayes Impact Fellowship is trying to transform these ideas into realities- with the help of big data, of course.

Bayes Impact launched their fall 2014 fellowship last week, with an impressive roster of talent to support the fellows’ grand ambitions. The founders alone are impressive- Paul Duan, formerly Eventbrite’s lead data scientist, Andrew Jiang, a non-profit veteran and Eric Liu, a former Thomvest Ventures analyst founded the fellowship earlier this year. The fellowship’s mentors and scientists hail from Airbnb, Ayasdi, Intuit, LinkedIn, Netflix, OpenTable, Salesforce.com, and Square- to name but a few.

Already, non-profits are leveraging big data to provide insights and spark change. We’ve previously reported on the use of big data in conflict prevention initiatives, and the Data Science for Social Good Fellowship, ran by Obama’s former chief data scientist. The Bayes Impact team want to see a greater scope of work in this field. “Almost all technology companies, no matter their size, use data to their and their users’ advantage: not-for-profits and governments can and should do the same thing,” Zachary Townsend, founder of the startup Standard Treasury and a Bayes Impact board member, told VentureBeat.

The fellowship lasts between 6 and 12 months, and the fall intake will be 20-25 fellows working on 10-12 projects. The Summer pilot project saw fellows tackling fraud prevention and assessing credit worthiness for the microloan initiative Zidisha, and tackling overcrowding in the Californian prison system. The aforementioned project to determine which buildings are most likely to catch on fire (using metrics from fire safety inspections) is already a confirmed project for the Fall programme.

“[The mentors at Bayes Impact are] pretty carefully selected, for the particular problems, not just somebody who has been working as a data scientist, but really knows about what problems in particular the fellow might be working on,” said summer fellow William Lane. Let’s hope the right fellows are teamed with the right experts, so that these lofty ideas can be translated into actionable change.

Read more here.
(Image credit: Bayes Impact website)

Follow @DataconomyMedia

Interested in more content like this? Sign up to our newsletter, and you wont miss a thing!

[mc4wp_form]