What it's about: This multi-year project addresses the information gap facing impact organizations by providing non-profit organizations with similar tools and industry analytics that for-profit companies have been using for years. The project is based on anonymized and aggregated Mastercard transaction data. underlying factors. By making the data available, these organizations can better understand trends in individual donations, as well as develop insights around various factors that effect giving.
Participants: Master Card, Center for Inclusive Growth
What it's about: This project used financial transaction data to understand how people behave before and after natural disasters by analyzing Point of Sale (POS) payment and ATM cash withdrawal data from more than 100,000 BBVA Bancomer clients during Hurricane Odile on the Mexican state of Baja California. Data analytics were employed to derive proxy indicators of the economic impact and market resilience of people in the region.
Participants: UN Global Pulse and BBVA Data & Analytics
ECONOMIC RESILIENCE TO NATURAL DISASTERS
What we like: The use of financial transaction data to improve response and understanding of the economic impact of climate change.
What it's about: Inferring the income level of individuals in locations is often important for public policy as it can inform the targeting and design of social programs and other policy areas. This study analyzed mobility signatures in debit card transaction records to infer the income level of people in major metropolitan areas of a country.
INCOME LEVELS & FINANCIAL TRANSACTION DATA
What we like: Using data and technology to provide impact organizations with private sector analytical tools to improve a crucial component of organizational success.
What it's about: The type of roof a house has is used as a proxy-indicator for poverty in this project. Traditional thatched roofs harbor pests and disease and are high maintenance and people upgrade to a metal of tiled roof when they are able to attain a level of wealth.
Participants: Uganda Bureau of Statistics, University of Edinburgh
COUNTING METAL ROOFS TO MEASURE WEALTH
What we like: Creative use of geo-spatial satellite photos data and technology to inform economic policy.
What it's about: This is an annual conference going into its third year that brings together researchers and practitioners interested in fairness, accountability, and transparency in artificial intelligence and socio-technical systems in general. The conference generates research and papers from a wide variety of disciplines, including computer science, statistics, the humanities, law, and education.
Participants: ACM and Multiple Sponsers
Conference on Fairness, Accountability,Transparency
What we like: Includes many of the leading thought leaders and ideas on ai and fairness
What it's about: This extensible open source toolkit's purpose is to examine and hopefully mitigate discrimination and bias in machine learning models. It containing over 70 fairness metrics and 10 bias mitigation algorithms developed by the research community, it is designed to translate algorithmic research from the lab into the actual practice of domains such as as finance, human capital management, healthcare, and education.
AI FAIRNESS 360 OPEN SOURCE TOOLKIT
What we like: Innovative Corporate Social Responsibility through open source AI Fairness tool.
What it's about: This ongoing project uses micro-finance data to understand which factors affect savings and loans mobilisation and whether micro savings and loans can offer insights on macroeconomic issues in Cambodia. Insights from this project could help inform policy on financial inclusion and early warning systems concerning economic shocks.
Participants: United Nations Capital Development Fund (UNCDF)
MICRO FINANCE DATA FOR ECONOMIC POLICY
What we like: Using financial data to inform and guide national economic understanding and policy.
What it's about: Google's machine learning diagnostic tool lets users try on five different types of fairness to evaluate different machine learning models as a means of making sure that AI doesn't perpetuate or exacerbate the unfairness of existing systems. The project raises the issue that, as a culture, we often lack consensus about which of the many types of fairness to apply and that determining fairness requires both technical and non-technical decisions and trade-offs.
Participants: Google Research - People and AI Fairness
PLAYING WITH AI FAIRNESS
What we like: Getting people to think about the harder questions of unfairness and bias in AI may help them see the problems that exist today.
What it's about: Lending to small entrepreneurs in most developing countries is an expensive and high-risk endeavor. Without credit bureaus to screen borrowers for creditworthiness, lenders must invest their own resources to assess each loan applicant’s ability to repay. This project helped to create a data-driven approach to lending and also investigate factors - including repayment behavior - related to loan default by building a predictive model that could be used to help determine default risk.
Participants: Microcred and DataKind
Financial Inclusion in Senegal Using Predictive Modeling
What we like: Demonstrates that developing a data-driven approach to lending can advance the mission of offering simple, accessible financial products and services to people who would not otherwise have access to the financial sector.
What it's about: While Libra is not a true block chain crypto currency, it is a massive project with huge participation from different organizations around the world. The combination of Facebook's data with what will likely be the most widely accepted crypto currency is simply a force that cannot be ignored. Without being naive, it is possible to feel some optimism about a system that potentially enables free peer-to-peer cross-border transactions for families who rely on remittances.
What we like: Demonstrates that change has only just begun.