add prepare_sector_split.R #30
Conversation
| abcd_id <- abcd %>% | ||
| dplyr::distinct(.data$company_id, .data$name_company) |
There was a problem hiding this comment.
NIT: single line pipe (i don't really care that much, just making note)
| # identify lost_companies_sector_split and write to csv for inspection | ||
| lost_companies_sector_split <- companies_sector_split %>% | ||
| dplyr::anti_join( | ||
| abcd_id, | ||
| by = c("company_id") | ||
| ) |
| config_files <- config::get("file_names") | ||
| config_match_prio <- config::get("match_prioritize") | ||
| config_prepare_sector_split <- config::get("sector_split") |
There was a problem hiding this comment.
Does this add new/necessary keys to the config? Should that be documented somewhere?
There was a problem hiding this comment.
no, they are already in the example.config.yml
It should be documented, but at this point I prefer getting the full workflow set up before writing an extensive documentation about these. But will open a task for it
There was a problem hiding this comment.
Sounds good.
We should consider how exactly we want to slice and dice the architecture here, and what/ where things should be documented (e.g. {config.yml + scripts} could also be {function args + functions} in a package).
The technical profile document will inform this, and ideally we can discuss it a bit more in the Tech Review.
closes #9
closes #10
workflow.multi.loanbookgains moduleprepare_sector_split.R, which calculates equal weights and primary energy based sector weights for companies"worst_case"sector split, which was an option in the previousworkflow.aggregate.loanbooks. This option has not been used and is unnecessarily complicated.workflow.aggregate.loanbooksrun_match_prioritize.Rgains an option to split loan values based on outputs ofprepare_sector_split.R. Option is set viaconfig.ymlprimary_energy_efficiencyandunit_conversion) needed to calculate the primary energy based sector split are integrated directly intoprepare_sector_split.R. Previously, they were generated indata-raw/, but since we are currently not using a package structure and the data sets are not expected to be used elsewhere, they are integrated directly in the script for now.prepare_sector_split.Rcan remove inactive companies based on the output ofprepare_abcd.R, in case the optionremove_inactive_companies == TRUE