This report presents a comprehensive quality control assessment of
the harmonized payroll data comprising the Contract,
Personnel, and Establishment modules
according to the standard harmonization dictionary
(govhr::harmonization_dict()). All diagnostic tables are
generated automatically using the compute_qualitycontrol()
function, which evaluates the aforementioned harmonized modules.
The quality control report is organized into four main sections:
Data Basics
Foundational diagnostics including module dimensions, variable structure
and dictionary conformity, primary key integrity, cross-module orphan
checks, salary logic validation, and date logic checks. This section
provides essential structural and consistency assessments before diving
into detailed data quality patterns.
Missingness Report
Comprehensive measurement and visualization of missing data patterns
overall and across key analytical dimensions (occupation, ISCO mapping,
reference period, and establishment). Identifies variables requiring
cleaning, imputation, or further validation.
Volatility Analysis
Time-series diagnostics evaluating temporal stability of salary
aggregates, contract counts, and work hours across reference periods.
High volatility may indicate data quality issues, organizational
changes, or policy shifts requiring investigation.
This section provides a high-level profile of the harmonized payroll data and summarizes the structure of the three core modules—Contract, Personnel, and Establishment—before detailed diagnostics are introduced.
The harmonized dataset includes the following components:
Module Dimensions Summary |
||
| Module | Observations | Variables |
|---|---|---|
| Contract | 8,885 | 23 |
| Personnel | 8,672 | 9 |
| Establishment | 47 | 9 |
This section visualizes data coverage across time periods, highlighting potential spikes or gaps in observations that may indicate data quality issues or structural changes.
This section compares the variables present in each harmonized module against the standard harmonization dictionary. It identifies:
Structure Diagnostics by Module |
||
| Module | Missing Variables1 | Extra Variables2 |
|---|---|---|
| Contract | personnel_id | worker_id |
| Establishment | est_name_en | est_name_english, ref_date, est_type |
| Personnel | birth_date, country_code | est_id, est_name_native, birth_day |
| 1 Red badges indicate variables expected by the dictionary but missing from the data | ||
| 2 Orange badges indicate variables present in the data but not in the dictionary | ||
This section validates that the Contract module’s primary keys
(contract_id, personnel_id,
ref_date) uniquely identify each record without
duplication.
Primary Key Uniqueness Check |
|
| Check | Status |
|---|---|
| Primary Keys Unique | ✓ PASS |
| Number of Duplicate Groups | 0 |
This section identifies orphaned records—IDs referenced in the Contract module that don’t exist in the Personnel or Establishment modules.
Orphan Records Summary |
||
| Check | Missing Count | Status |
|---|---|---|
| Personnel IDs in Contract missing from Personnel module | 0 | ✓ PASS |
| Establishment IDs in Contract missing from Establishment module | 0 | ✓ PASS |
This section examines patterns of missing data across the Contract,
Personnel, and Establishment modules.
Understanding missingness is essential for data quality control because
it highlights variables that may require cleaning, imputation, or
further validation by the data provider.
We evaluate:
This section evaluates temporal stability across key HR metrics—salary aggregates, wagebill, staff counts, and contract counts. Volatility is measured using rolling coefficients of variation (CV) or percent changes to identify establishments with unstable patterns over time.
High volatility may indicate:
The heatmaps below show volatility metrics for the top 50 contracts with the highest average salary volatility, allowing identification of patterns that warrant further investigation.
