Random Hacks of Kindness at Oxford – 40,000 records test data transparency

Last month I joined an unusual group of techies gathered at Oxford for an Oxfam sponsored event that allowed us to get cracking at code to help with social problems like flood forecasting, fair-trade, and international aid transparency.  Random Hacks of Kindness (RHoK) sponsored this and more than 30 other events on December 3 (our Fundacja TechSoup colleagues organized the parallel event in Warsaw) where specific social problems were matched to a group of volunteer geeks to see if technology can help.

Founded in 2009, RHoK is the vision of a brain trust from Google, Microsoft, Yahoo!, NASA and the World Bank to create a global community of innovators building practical open technology solutions to help change the world.  On the RHoK “solutions” page you can see how this works: terrific hackathons to create apps to help with disaster relief (matching families and victims), sharing medical records, and mapping everything from potential earthquake faults to financial defaults in poor neighborhoods.

For example, at our Oxford event there was a project undertaken by Stewart Marshall and his team to create a map to visualize data on humanitarian projects so that an aid agency can better tell its story to supporters and donors. They used data on development projects from the International Aid Transparency Initiative (IATI).

As for me, since TechSoup Global has a deep interest and engagement through its services in projects involving assessment of data from NGOs in many countries -- I suggested that we think of an app that could evaluate the accuracy and relevance of the NGO data that’s already out there posted by myriad public and private agencies.  The challenge is to (a) make sure the data is accurate and (b) make sure the data is in a useful format that philanthropists need.

The solution I proposed was to create rules of inference using heuristics, predefined rules and machine learning to assign scores to each NGO. Together with Dirk Gorissen, Chris Wallace, Ben Foxall, and Russell Law, we tested it on a sample dataset that I created from the NGO Partnerships website of the Indian Planning commission – 40,000 records!

Would you like to see our project? Here’s a walkthrough of our ‘NGO Clarity’ idea! 

(You can read about other RHoK posts on the TSG blog here and here.)