Questions for Jay Qi, Data Scientist at DrivenData
In the realm of data science, DrivenData has emerged as a leading force in harnessing the power of crowd-sourced machine learning solutions to tackle complex social issues. Over the past eight years, the organisation has run over 65 competitions, awarding over $3.3 million in prizes, and making a significant impact in various sectors such as health, energy, agriculture, and public safety.
One of DrivenData's most impactful projects includes the development of a real-time, interactive map to visualise attacks on the Ukrainian healthcare system. This tool supports humanitarian aid efforts, legal accountability, and raises public awareness of the crisis. Another notable project is the collaboration with The World Bank to analyse repayment data from pay-as-you-go (PAYG) off-grid solar companies, helping to build key performance indicators that can attract investment and improve energy access for over 1 billion people lacking electricity.
DrivenData's projects extend beyond data analysis. They have also created interactive visualizations for Fair Trade USA to trace product origins through the supply chain to retail destinations, enhancing transparency and supporting ethical sourcing. In Tanzania, they analysed millions of mobile financial transactions to identify behavioural patterns and trust barriers, guiding the design of user-centric mobile money services to boost adoption among low-income populations.
The organisation also partners with Yelp, Harvard, and the City of Boston on a predictive challenge linking Yelp reviews and ratings to food safety inspection results, helping improve public health monitoring. Disaster resilience in urban areas is another focus, with challenges like the “Open Cities AI Challenge” tasking participants with segmenting buildings in aerial imagery to support disaster resilience planning.
Ethical AI, transparency, and collaboration with mission-driven organisations like NASA and The World Bank are integral to DrivenData's approach. They also consult with such organisations, maintain open-source software tools, and publish learning resources.
DrivenData's online machine learning competitions attract data scientists from around the globe. Winning models' code and documentation are required to be open-sourced, ensuring transparency and enabling further research and development. The competitions cover a wide range of application areas, including sustainability, health, social media moderation, and more.
Open data released after competitions serves as a valuable resource for future research and development. DrivenData has also participated in advancing research in privacy-enhancing technologies through competitions in partnership with NIST and other agencies.
Effective use of data requires investment in technology, processes, and staff. DrivenData employs human-centered design principles in their data science consulting work, collaborating with partner organisations to understand their needs and identify the right approach to a problem.
In summary, DrivenData's competitions have generated data-driven tools and models that enhance humanitarian aid, sustainable development, public health, financial inclusion, and disaster preparedness worldwide. By leveraging the collective intelligence of the data science community, DrivenData continues to make a significant impact in addressing global challenges.
- DrivenData utilizes artificial intelligence and machine learning to power crowd-sourced solutions, impacting social issues in sectors like health and energy.
- One of DrivenData's notable projects was developing a real-time map for the Ukrainian healthcare system, supporting humanitarian aid, accountability, and public awareness.
- In collaboration with The World Bank, DrivenData analyzed PAYG off-grid solar repayment data, aiming to improve energy access for over 1 billion people.
- DrivenData's projects extend beyond analysis, creating visualizations for Fair Trade USA to enhance supply chain transparency and support ethical sourcing.
- In Tanzania, they analyzed mobile financial transactions to identify behavioral patterns and trust barriers for user-centric mobile money service design.
- DrivenData's projects also encompass predictive challenges, such as linking Yelp reviews to food safety inspection results for improved public health monitoring.
- Encouraging ethical AI practices, transparency, and collaboration, DrivenData partners with organizations like NASA and The World Bank while maintaining open-source software tools and publishing learning resources.
- The winning models' code and documentation from DrivenData's competitions are open-sourced, enabling further research, development, and innovation in data science.
