Last month, we celebrated nearly a decade of harnessing the power of data science and AI in the service of humanity. It seems like just yesterday that we hosted our first-ever DataDive® event. Since that weekend in New York, so much has changed, but a few things remain the same. Our volunteers are the heart of DataKind. Our project partners are at the frontlines of social change. And the world needs DataKind now more than ever.
Since our founding in 2012, we’ve fostered an international community of 20,000+ skilled technologists. With roots in New York City, we collaborate with Chapters in Bengaluru, Singapore, San Francisco, United Kingdom, and Washington DC. This collective force successfully completed 350+ projects and delivered more than $35 million in pro bono services. But the demand for our data science and AI expertise is great, and there’s still so much more to be done.
The scale and unpredictability of global challenges to come will only expand over time. The time is now for us to mobilize even more people who are driven to solve problems with social impact, and we’re excited to help shape the use of data science and AI so they serve all of humanity. We can’t wait to see what Year 10 will bring. To our past and present team members, volunteers, project partners, and funders, thanks for dedicating your time, talents, and treasure to building DataKind’s global community and expanding our work around the world.
In honor of nine years, join us in celebrating nine of our project partners (in no particular order!) and learning how we helped to accelerate the pace of their impact.
Increasing access to sanitation
Waste management and sanitation are significant challenges in many countries. Approximately 36% of the global population lack access to a toilet, contributing to health and environmental issues. SOIL is a nonprofit that addresses this lack of dignified sanitation by providing toilets to communities and then removing waste so it can be converted into fertile compost. There were two challenges to SOIL’s sanitation services: transportation costs and logistics and lack of easy-to-use optimization tools. To address these problems, we partnered with SOIL to create software to deliver their services more efficiently, leading to increased community access to mobile sanitation amidst COVID-19, improved transportation efficiency by 10%, and reduced costs and fuel use.
Improving college success
Nearly half of all students entering college in the U.S. are at risk of leaving without earning a degree. John Jay College, located in New York City, inquired how existing student data and machine learning techniques could be combined to address challenges surrounding graduation completion rates. We analyzed more than 10 years of historical student data and built a predictive model that was used to design effective interventions to increase graduation success. The software and insights helped John Jay administration identify students challenged to reach graduation so the university could intervene sooner and, ultimately, increased the graduation rate to 73% (from a projected 54%) in the first year.
Preventing home fire deaths
Many fires are preventable and often occur as a result of limited information, including smoke alarm testing and fire escape planning. In response, the American Red Cross (Red Cross) runs information campaigns to prevent fires. A challenge it faces, however, is identifying which communities are most at risk in order to properly allocate their resources and provide targeted information campaigns. Red Cross joined forces with a team of DataKind DC volunteers and built the Home Fire Risk Map tool to identify neighborhoods at risk of fires to target for in-home fire safety education and smoke alarm installations. Since the launch of the campaign in October 2014, the Red Cross, in partnership with local fire departments, conducted nearly 900,000 home fire safety visits, installed more than 2.1 million smoke alarms in communities that need them most, and documented 794 lives saved.
Optimizing access to healthcare
Most frontline health organizations rely on paper-based record-keeping. However, it can take weeks or even months for data on paper to be digitized. Riders for Health (Riders) expands healthcare to 47 million people across five African countries, many of whom live in rural and remote regions. A key service of Riders is the transport of medical samples for diagnosis (including tuberculosis, HIV, and COVID-19). To maintain the quality and integrity of samples and results during collection, transportation, storage, and analysis, the Riders’ couriers capture details in handwritten logbooks. This time-consuming process requires the courier to manually copy information about the medical samples they’re transporting into their logbooks. While important, completing paperwork limits the volume and frequency of visits a courier can make between health facilities and patients. We used computer vision techniques and machine learning to demonstrate the time currently taken to digitize paper-based records can be reduced from nearly 60 days to under a day.
Reducing case worker burnout
In the world of foster care, caseworkers play a critical role in making sure that children find permanent, safe homes. A caseworker is often a child’s main point of contact and source of security, so it can be extremely disruptive if children have to switch caseworkers. The caseworkers at Embrace Families of Central Florida, a nonprofit that serves over 3,000 children, were overburdened and experienced massive scheduling and logistical challenges. Working with the organization, we helped to lessen the risk of burnout and increase the likelihood of permanency for the children by creating a scheduling algorithm that reduced travel schedules by as much as 30 miles/day, saved $2,025/day in travel costs, and added hours for one-to-one foster care time among the 150+ caseworkers.
Increasing access to water
Moulton Niguel Water District (MNWD) provides water, recycled water, and wastewater service to approximately 170,000 people across several cities in Southern California where droughts have plagued local communities for the past several years. Facing the worst drought California had seen in the past 500 years, MNWD needed to more precisely and accurately forecast water demand to improve pricing, expand conservation programs, and increase the efficiency of water transportation throughout its network. With this goal, we created data infrastructure and dashboards to predict customer demand more accurately and saved the agency an estimated $25 million in avoided capital storage costs.
Accessing electricity in rural households
Approximately 980 million people around the world don’t have access to electricity. Simpa Networks is a technology company with a bold mission: to make modern energy simple, affordable, and accessible for everyone. To help Simpa Networks make its energy system available to more people in India, we used historical customer data to build a model which predicts which applicants are a better fit for the payment assistance program, giving more accurate projections of loan repayment. This model allowed Simpa to better plan their cash flow and growth, while offering opportunities to a majority of applicants so that more individuals across the community have access to affordable electricity.
Tackling online hate speech
The Southern Poverty Law Center (SPLC), a legal advocacy organization specializing in civil rights and public interest litigation, partnered with us to reduce online hate speech. SPLC tracks more than 1,600 extremist groups operating across the country. It publishes investigative reports, trains law enforcement officers, and shares key intelligence and expert analysis with the media and public. We developed an algorithm to monitor how effective hate sites are at exploiting search engine marketing and optimization so SPLC can exert more pressure on search engine companies to take more responsibility over their algorithms. And by automating these processes, SPLC reallocated staff time towards understanding how hate speech appears in web searches and how to reduce its mainstream exposure.
Protecting wildlife worldwide
The Tropical Ecology Assessment and Monitoring (TEAM) Network is the largest global-scale conservation observation network, functioning as an early warning system for many species. Their Wildlife Picture Index tool uses a multitude of data points to monitor the health and status of over 500 species populations around the world. Conservation International (CI), a nonprofit organization that embraces innovation to protect plant and animal diversity, uses the Wildlife Picture Index to inform its work and wanted to explore how to improve it. A team of volunteers from DataKind SF worked with CI to develop an interactive heat map to analyze trends within specific species at specific sites, ensuring land use managers, wildlife ecologists, and other decision makers get the information they need to better protect wildlife.