As suggested by our name, we are putting as much knowledge, learning and data out in the open as we can. We believe that no one can mitigate climate change alone, and through curating our code as a usable open source project, our work stands to have its greatest impact and widest reach.
The climate emergency is waiting for no one. To allow us and all the world to work globally and at speed, we aim to:
As we work to establish an impactful open source project, we perform all our development completely in the open. Our code, GitHub issues and plans are all published and documented on github.com/openclimatefix. Projects range from our DataPipes to implementations of well-known academic papers from Google Research and Google Brain, to our own forecasting techniques.
We want people to use our code or have a go at tackling a few GitHub issues. If you’re wondering where to start, here is a list of good first issues!
Data is the lifeblood of machine learning. And energy data is often difficult to obtain, messy to work with and not easily shared. At Open Climate Fix we are sharing as much data as we are legally able to. In many cases these are extremely large datasets not usually available to the public. We have datasets on numerical weather predictions (NWP), satellite imagery and PV time series spanning the globe in many cases.
We try to document this data as well as possible, and have put considerable effort into processing the data into a format that makes it as usable as possible.
Please visit our data on huggingface.co/openclimatefix to see the data we’ve published. On Hugging Face we have 10s of terabytes of NWP data, 8 years of satellite imagery at 5-minute intervals covering all of Europe and some of Africa and the Middle East as well as multi-year datasets of precipitation radar data covering the UK, Sweden, and the US.
You can also contact us at email@example.com.