About Quant Climate

Quant Climate uses Bayesian machine learning to quantify the risks of a changing climate and the effects they have on the economy. We work with problems related to finance, insurance, and the public sector where it’s crucial to understand the impacts of extreme weather.

What we do

We build models that offer accurate point estimates along with reliable uncertainty quantification. That might mean forecasting how warmer summers will raise insurance claims, measuring what a clean-air policy changed once the weather is taken out, or estimating how exposed a portfolio is to extreme heat. In each case you get the full range of plausible outcomes, so you can plan for the bad end as well as the middle.

We also build the tools we use and publish them as open source, so anyone can see how they work:

• GPJax, a Gaussian process library for JAX. We use it to map how temperature and rainfall vary across space and time, including where measurements are sparse.
• Impulso, for Bayesian vector autoregressions in PyMC. We use it to trace how a shock such as a drought works through prices and output over the months that follow.

Tell us what you need to measure. Get in touch.