Even people who do not believe in its existence seem concerned about climate change. While debate, data, and politics spin around the issue, decisions on what to do remain uncoordinated. Great scientists work on modeling the climate’s reaction to human forcing, and great policy experts debate the right ways to mitigate the ill effects of humans’ actions. The data and models on which climate researchers and policy makers rely all rest on tricky-to-measure combinations of physical and human processes, and it is easy to see why modeling, and planning for, the combined effects is so challenging. The inputs to the models and their predictions are uncertain, and interactions amongst key processes make the challenge of quantifying the uncertainty even harder. Ultimately, though, we need to know which decisions will have the greatest effect, in what way, when, where and how. Data science approaches, where numerical and statistical tools are brought to bear on vast quantities of diverse quantitative information keep giving ever-more-impressive results to ever-more-difficult problems. That said, we do not believe in “data science” as a panacea. But, we do wonder how its methods might be more effectively used in specific areas of climate change research and in connecting aspects of climate change modeling and policy outcomes that are not typically considered together. Only recently have a few groups come together to even talk about how best to combine climate and data science research more intimately, and no one has yet made the great, integrative, strides that should be possible in a coordinated program. So, with the Accelerator Workshop proposed here, we aim to start a more focused effort, beginning with conversation about if and how it is possible to use new data science approaches to slow down and mitigate climate change. The physical and human systems we need to understand to correctly model the world’s possible futures are many, and the network of interconnections amongst those systems is vast. Imagine a roomful of experts representing: architecture; biology; chemistry; computer science; earth science; economics; environmental science; epidemiology; food science; history; hydrology; mechanical engineering; linguistics; medicine; political science; physics; religion; sociology; solar physics; and statistics. Now imagine a group of researchers who run global simulations (of weather, geography, politics, wealth, health, and combinations thereof) entering that same room. Add a group of data scientists, eager for challenging questions about complicated interactions amongst quantifiable systems. Fruitful conversations, ideas, and collaborations generated in this hypothetical room could hold the keys, literally, to saving the world. The long-term goal of our proposed project is to inspire and support transdisciplinary research on climate generated by these conversations, using data and data science.