UNDP Accelerator Labs
Towards
mainstreaming
data innovation
for sustainable
development

Over the past four years, the UNDP Accelerator Labs Network has been pushing the mainstreaming of data innovation* inside and outside UNDP. Here, we take stock of these efforts, focusing specifically on work that leverages existing, third-party data, and includes some form of advanced computational technique—as opposed to work that focuses on more manual means of inputting and analyzing "new" data, like Collective Intelligence. The work is action oriented, rather than (solely) research oriented, meaning it is not focused on pushing the limits of data innovation, but rather on its integration into practice.

The interactive graphic below highlights exemplar applications from around the world. It is a living document, meaning we plan to update it regularly.

*We refer to data innovation as the use of unusual data sources and advanced computational techniques for sustainable development practice. We prefer the terms "unusual data sources" and "advanced computational techniques" over "big (or new) data" and "Artificial Intelligence (AI)", as we believe that the prior more adequately reflect the situated notions of resource and tool, while the latter are, in our opinion, generally confounded by the current hype around Generative AI. We see data innovation as a means to an end, not an end in itself.

The Accelerator Labs “work out loud”, meaning they continuously publish updates on their work, whether through blogs or action learning plans and reflections. Here, we sample 29 projects from 21 different Labs using these sources of information, looking for terms like “data innovation”, “unusual data”, “big data”, “data science”, “machine learning”, “artificial intelligence”, and “AI”.

We code each project as a geographically situated triptych of data source—computational technique—application area. Our intention is to uncover convergences that might suggest maturity of given data innovation techniques. Note that we use relatively high-level descriptive classes for our coding to facilitate aggregation. For example, we characterize computational techniques as “object detection” or “text classification”, rather than by specific types of classification or of machine learning models.

Read more about this work in our blog post: Dismantling the AI Monolith for Sustainable Development – Part 1: Observations on Our Use of Data and Computing

Find all source materials on Github.