carbonfirst / CarbonCast

A system to predict hourly carbon intensity in the electrical grids using machine learning. CarbonCast provides average carbon intensity forecasts for up to 96 hours.
Apache License 2.0
33 stars 8 forks source link

Add data for Taiwan #5

Open fershad opened 9 months ago

fershad commented 9 months ago

Taiwan has a bunch of open data around grid mix which, I think, could be added to this project. It would be great to have an Asian grid represented, as it might hopefully lead the way for others.

I've linked to some of the data below. I'm not sure if this is in the format required for this project, but would be happy to help with translations so that it can be better understood. I'm not proficient with Python, but with some guidance I might be able to help with a PR if there's a desire for that.

Open historical data

Forecasts

If there's other data that would be handy, let me know & I'll see if I can track it down.

diptyaroop commented 9 months ago

Hello, Yes, it would be great to have an Asian grid represented.

We know about the ElectricityMaps data, but the problem is that we cannot provide real-time forecasts using ElectricityMaps data since it is not available in real-time (for example, they have not released 2023 data yet).

What we can do is get the data from other sources (for example, Taipower, as you have listed), & tally with ElectricityMaps for validity. The TaiPower source looks promising, and we will look into it in more detail. Thanks for sharing. However, one thing we noticed is that the temporal granularity is in days & not hours. If you can find an hourly electricity data source for Taiwan, that would be highly beneficial.

Also, if you are interested, you can write a script, gather the data (refer here for data files/format required), and raise a PR. That would be very helpful. Maybe this file will help you in getting started with writing such a script. Please also feel free to ask any questions/doubts here as well.