[x] [P1] Select a bounding box from the map, not just countries
[ ] [P2] Select aggregated countries, e.g. EU --> ask James about GIS libraries or datasets that might help
[x] [P1] If we keep using Natural Earth Data, we should make it more flexible. Users should be able to select either countries including their oversea territories, or map units. We can take inspiration from the Wildfire Emission Explorer, as it was pretty flexible and had all they needed
[ ] [P3] Change Folium map projection to account for surface area
[x] [P1] Unit conversion for GHG fluxes: we are interested in the yearly total, not average
[x] [P2] Unit conversion for GHG: multiply flux per unit area to each country area
[ ] [P4] Calculate pixel fraction for GHG data vs. country polygon
[ ] [P4] Make it flexible to calculate or get the confidence interval for the GHG time series plot
[x] [P1] Flexible selection for GHG flux variables (e.g. vegetation is actually more important than fossile fuels)
[ ] [P4] Think about combining data from different datasets
[ ] [P3?] Caching of data in Streamlit application
[P1]: definitely must do
[P2]: slightly less important than P1, but still a must-do
[P3]: nice to have, smaller detail
[P4]: we shouldn't aim to do it, but just keep it in mind and document how it should be done
[x] GHG variables and units mapping: see also related TODOs above (and notes from 17/7 call, pasted below). We are only interested in flux variables and mostly in some of them (e.g. a posteriori is more important than a priori)
[x] Allow users to select data variables from Streamlit app
[P1]: definitely must do [P2]: slightly less important than P1, but still a must-do [P3]: nice to have, smaller detail [P4]: we shouldn't aim to do it, but just keep it in mind and document how it should be done