Precision of emission factors is misleading about electricitymaps-contrib HOT 5 OPEN

But how should we handle it when some of the scientific sources we use have higher precision?

Good point - I think this could be an exception.

from electricitymaps-contrib.

I don't think these need to be rounded in the config file if they're just intermediate results from some calculation like the ETS/ENTSO-E mapping. Otherwise, you could argue that before starting the emission factor calculation, total emissions and total electricity generation of each plant should be rounded as well, and then round those intermediate emission factors of power plants before combining them for the country factor. So IMHO, intermediate values should be stored in high precision. (Of course we don't really need 7 decimal places, but I wouldn't say that's something that needs fixing either.)

When emission factors are displayed in the app / website for public consumption though, they should always be rounded in some way of course, to avoid the misleading impression. I believe that's already implemented everywhere?

from electricitymaps-contrib.

I don't think these need to be rounded in the config file if they're just intermediate results from some calculation like the ETS/ENTSO-E mapping. Otherwise, you could argue that before starting the emission factor calculation, total emissions and total electricity generation of each plant should be rounded as well, and then round those intermediate emission factors of power plants before combining them for the country factor. So IMHO, intermediate values should be stored in high precision. (Of course we don't really need 7 decimal places, but I wouldn't say that's something that needs fixing either.)

When emission factors are displayed in the app / website for public consumption though, they should always be rounded in some way of course, to avoid the misleading impression. I believe that's already implemented everywhere?

Thanks for your view and insights, and yes they are already rounded in the frontend.

And I suppose an argument could be made that these values are actually highly accurate (not accounting for floating point errors) but are missing some source data (not all plants have emissions reported and vice versa) which makes the derived value "wrong" but accurate.

from electricitymaps-contrib.

I think rounding to the whole number makes the most sense, at least for our calculated numbers, otherwise things might be even weirder since nuclear currently is below 10g for most of Europe.

But how should we handle it when some of the scientific sources we use have higher precision?

from electricitymaps-contrib.