[ ] Performance improvements. In combination, these would drop run time from ~4 minutes to tens of seconds. This is a qualitative step change from "run it and get a drink" to an interactive process. Interactivity allows for more experimentation and error checking due to the reduced costs of experiments. These are also small changes; probably 1.5 hours total. Low cost, medium benefit.
[ ] cache pudl MCOE calculation. It takes over half the total run time and never changes.
[ ] update db writes to use COPY instead of shockingly slow INSERT. I've already implemented this in another project, just have to copy it over.
[ ] #206 . Medium cost, medium benefit.
[ ] move PAD-US analysis inside repo instead of storing results as a flat file. The only reason it is arranged as such is that we didn't have a cloud data cache set up at the time and the input data was too large for git. Low cost, low benefit.
[ ] #166 Medium cost, low benefit.
[ ] move Energy Communities analysis inside repo, if the dependencies can be reasonably resolved. I suspect this project has been abandoned by RMI, so I don't expect to lose access to future updates. Medium cost, low benefit.