DataDives are high energy, marathon-style, 48-hour events where nonprofits work alongside teams of data scientists, developers and designers to use data to gain insight into their programs, address key problems in the communities they serve and advance their missions. These inspiring community events welcome 50-150 data scientists and technologists to collaborate and together analyze and visualize data sets to provide partner organizations initial insights or prototypes to inform their work and create real world change.
Using data for good
Software, sensors and mobile phones produce a goldmine of data. Today, most companies are working closely with data scientists to get the most out of this data to better serve their customers. Datakind works with non-profits to uncover the power of data science and predictive analytics for their mission-driven organizations. Through a DataDive or two-day hackathon-style event, Datakind brings together volunteer data scientists and social change organizations to solve key problems in their communities.
A group of ‘Data Ambassadors’ or experienced volunteers work closely with the nonprofits during the six weeks leading up to DataDive to identify the most salient data questions, and to prepare or ‘clean’ the datasets for the DataDive. Participating in a DataDive serves as a unique opportunity for volunteers to learn and test out new skill sets, build cross-industry connections, and utilize their data science talents to produce actionable, meaningful results for mission-driven organizations.
The latest DataDive took place in July 2019, hosted by the Elsevier office in London. We also collaborated on a DataDive in London in November 2018, as well as in New York in July 2018 (here and here) and in London in November 2017 (here and here).
Machine Learning and Disaster Risk Management
In 2018, DataKind and the Elsevier Foundation have partnered on a DataCorps Disaster Relief project with the World Bank’s Global Facility for Disaster Reduction and Recovery (GFDRR) in support of the Sustainable Development Goals. The project analyzed high-resolution satellite imagery to create a scalable algorithm to distinguish among different building structures in order to identify rural areas susceptible to large-scale natural disasters. It aims to aid disaster risk reduction and recovery efforts in various developing countries.
The project was inspired by a one that DataKind previously piloted in rural Kenya and Uganda. Here, DataKind was able to develop a deep learning algorithm using satellite imagery to identify thatch and metal roofed homes as a way to determine which villages were most in need, given that metal roofed homes tend to signal higher household incomes. While the algorithm ultimately was not implemented, the project demonstrates the potential for nontraditional data sources like satellite imagery and cutting edge techniques like machine learning to help fuel the work of social change organizations by helping them automate time-consuming processes, understand the communities they serve and ultimately reach more people in need.
Beneficiary group: The 2019 DataDive worked with 3 non-profits:
- Street League supports unemployed 16-24 years old to move into employment using the power of sport, teaching them key life and work skills they need for sustainable employment.
- Mind is driven by the needs of people experiencing mental distress by offering a wide range of services and support in London.
- The Mix provides a free, confidential multi-channel service offering support for under 25 year olds – so that all young people are be able to make informed choices about their physical and mental wellbeing, to live better lives.