Harmonization Assembly Line

deposit_replication_package_to_bucket.py

postgres · gprl-harmonization

replication_package

replication_package_file

metadata_field_definition

metadata_assertion

replication_package_file_metadata

prepared packet

packet · file_type · baseline_hh.dta

prepare_study_metadata_extraction_packets.py

upsert_assertion_log_into_database.py

scan_and_catalogue_replication_package_contents.py

study_core_panel_construction_scripts/S_4k7qb_kar2019.py

construct_core_panel_from_raw_tables(raw)

unit hhid from baseline_hh.dta arm treatment_arm from study_arm occasion wave from study_measurement_occasion keys only · no measurement columns

harmonization · the recorded decision

question → dta_var → universal_subcat → harmonized_var

op_raw_to_dta what the author’s code didagent op_dta_to_universal clean/merge into the conceptagent op_universal_to_harmonized the expression, evaluated by codescript

harmonization

assembled harmonized panel

build_harmonized_microdata/ · pure library

convert whose_value → household · recall → monthly · currency → PPP evaluate the parsed expression, by code assemble stack every study’s sealed part → one panel + manifest no gate · no model · no database write

Ingestion