Climate change is an issue we can’t solve alone, yet it seems like very few of those working to solve it have a big picture view of who is already working on what or where their work fits into the broader solutions landscape. Without a collaborative approach, effectiveness of private philanthropic and impact investment capital is severely limited. Climate funders need transparency, effective tools, and a holistic view of the landscape in order to make better decisions to set high-impact priorities.

The Climate Finance Tracker (CFT) is a suite of visual interfaces engineered by Vibrant Data Labs (VDL), an organization housed in Berkeley, California, that combines data and network theory into flexible tools, to tackle systemic social challenges like climate change. The CFT visualizes climate funding flows to organizations and companies on the ground. What started as a simple United States finance tracker is now poised to scale into Europe, Africa, and Latin America.

We recently caught up with Eric Berlow, founder of Vibrant Data Labs and co-creator of the CFT, who was awarded an Emerson Collective Climate Fellowship. He shared more about his background, philosophy, and goals.

To identify those gaps, you’d have to have tags and categories in the first place, correct? How does that work?

Eric: We currently start with philanthropy and investment data from Candid and Crunchbase (with more on the way!). We then gather, from online sources, more data on how the grantees and investees describe their work. This allows us then — using various methods, including natural language processing and machine learning — to categorize the organizations and let them self-organize into themes — all based around who is working on similar things.

A key challenge has been to develop a method for searching for ‘climate relevant’ investments and grants.  To do that we start with broad topic searches — for things like ‘climate’ and ‘agriculture,’ but then we need to filter these results because not all agricultural solutions are climate-positive, or some may mention ‘climate’ but in the wrong context. To do that we manually review a random subset of the results and use that to ‘fine-tune’ a Large Language Model to identify in the remaining results which are actually relevant to climate — for example, companies that are addressing things like regeneration, soil health, and sustainable water usage.

Read the full article about the Climate Finance Tracker by Peter Tavernise at Cisco.