Data for Good: As two worlds collide
When I lived in San Francisco, I used to be a member of the local chapter of DataKind (www.datakind.org). DataKind is an organisation that brings together data scientists typically from the private corporate sector to tackle data problems from nonprofit organisations. You would have engineers from the likes of Google and Uber tackling problems from homelessness in the city to landmines in Cambodia.
After many years working in international non-profits, I could relate all too well to the non-profits who presented their problems and their data. As a wannabe data nerd myself, I could also appreciate the revolutionary nature of the solutions that the data engineers would develop for their non-profit counterparts.
I loved going to these meetings just to see the meeting of the two worlds that I always thought needed to collide a little more often. They each came with their own realities and their own toolkits from which to draw.
There is a lot of interest in bringing the data revolution to the non-profit and public sector (See UN Global Pulse Lab, the Global Partnership for Sustainable Development Data, FIBR, insight2impact, World Bank Open Data, ITU’s AI for Good Global Summit, and countless open government data initiatives around the world. Kaggle has this year launched its second data for good challenge and DrivenData runs data science competitions for good). However, the revolution has been somewhat slow to take hold.
I see three fundamental challenges to translating the techniques and tools that the private sector has developed so well for profit, to non-profit development problems:
Data scientists and non-profit actors live in different worlds: The two sides rarely have a good understanding of each other, from concerns and priorities to rhythms of work and funding, not to mention the actual product of what they do. They don’t even speak the same language. While the data scientist is talking “random forest” and “neural nets” the non-profits are talking “theory of change” and “beneficiaries.” (I have been in these actual conversations.)
Data poor: For as much as the world has become data rich on a whole, large swaths of the population around the world still live in the data dark. For the majority of projects, whether on homeless youth or micro-finance for enterprises in the informal sector, there is rarely a nice clean reliable set of data. People are hard to locate, names change, “data” are of hand-written forms stored in a file cabinet in an office with no internet connection. Non-profits often face most of their issues around data collection. Forget data modelling.
Monitoring and Evaluation (M&E): If you ask a non-profit about their biggest data challenges, their mind will usually go automatically to M&E of their program. For non-profit organisations, M&E is their gold. It is what their donors require to fund them. What impact is attributable back to their specific program? AI and machine learning are the exciting developments coming from the industry. It was born out of the need to lower costs and improve marketing. It is not always so clear how to translate and demonstrate this power of data to non-profits’ needs.
Do social problems not match the data skill set coming out of the industry? I believe they do. But to have impact, data scientists working with development organisations need a keen understanding of the problems and an appreciation for the hard data realities of working on the ground. For development practitioners, they often have a hard time getting their heads around the opportunities of data analytics and how it applies to them.
There are incredible opportunities to unleash the power of data science for good. Attending those DataKind meetings definitely gave me hope of the catalytic power of bringing together data scientists and the non-profiters into the same room to talk.
At Ixio, we are also trying to tackle some of Africa’s most pressing social problems. We have a diverse team with the rare combination of both a deep understanding of development issues and the realities of the developing market context as well as the latest advances in machines learning and artificial intelligence. With this background, we are able to translate the skills that we have honed in other sectors to truly speak to the social challenges that the development sector faces.
Stay tuned as we launch a number of new financial inclusion and data for good projects and please reach out to us if you are interested in being part of the data for good revolution.