Compute Comprehensive Quality Control Checks for Harmonized HRMIS Data

Runs a full suite of QC diagnostics across the three harmonized HRMIS modules—contract, personnel, and establishment. Rule-based validation uses the validate framework and the built-in govhr::contract_rules / govhr::personnel_rules datasets, optionally extended with country-specific rules supplied via custom_rules.

Usage

compute_qualitycontrol(
  contract_dt,
  personnel_dt,
  est_dt,
  custom_rules = list(contract = NULL, personnel = NULL)
)

Arguments

contract_dt: A data.frame or data.table — harmonized Contract module.
personnel_dt: A data.frame or data.table — harmonized Personnel module.
est_dt: A data.frame or data.table — harmonized Establishment module.
custom_rules: A named list with optional elements contract and personnel, each a data.frame with the same four-column schema as govhr::contract_rules (rule, name, description, label). These rows are appended to the package-level rules before validation. Use NULL (the default) to run only the built-in rules.

Value

A named list:

n_obs: Row count of the Contract module.
n_vars: Column count of the Contract module.
structure: Dictionary-comparison results for each module.
orphans: Orphan ID diagnostics.
validation: Named list: contract and personnel audit reports from validate_data().
missingness: Long-format data.table of missingness by group, with columns group_var, group_label, group_val, target_var, target_label, n_missing, N, and pct_missing.
volatility: Named list: contract and personnel flat data.tables from flatten_volatility().
temporal_coverage: Named list: contract, personnel (if supplied), and establishment (if supplied), each a data.table with columns ref_date and n_obs showing the observation count per snapshot.
metadata: Named list with date_range (min/max ref_date across all modules), n_obs (row counts per module), and n_vars (column counts per module).

Details

Structure checks: Variable names checked against the harmonization dictionary for each module.
Orphan checks: Cross-module referential integrity: personnel and establishment IDs in the Contract module are checked against their parent modules.
Rule-based validation: govhr::contract_rules and govhr::personnel_rules (plus any custom_rules) are evaluated via validate_data(), covering uniqueness, date logic, salary consistency, working hours, age ranges, and more.
Missingness profiling: Missingness by key grouping variables (contract type, occupation, reference date). Groups are silently skipped if the column is absent from the data.
Volatility analysis: Period-over-period percent change for salaries, contract counts, working hours, occupation diversity, and contract-type diversity, each by multiple groupings. Groupings are silently skipped if the required columns are absent.

Examples

if (FALSE) { # \dontrun{
qc <- compute_qualitycontrol(
  contract_dt  = bra_hrmis_contract,
  personnel_dt = bra_hrmis_personnel,
  est_dt       = bra_hrmis_est
)
qc$validation$contract
qc$volatility$contract
} # }