Analyze Employment Overlaps Using over_id — analyze_employment

Provides comprehensive analysis of overlapping employment periods using the over_id functionality from vecshift() output. This function specifically focuses on periods with multiple simultaneous employment contracts (arco > 1) to understand patterns of multiple job holding, employment intensity, and temporal overlap characteristics.

Usage

analyze_employment_overlaps(
  segments,
  analysis_level = "person",
  min_overlap_duration = 1,
  include_temporal_patterns = TRUE,
  group_by = NULL,
  temporal_aggregation = "month",
  include_unemployment_context = TRUE
)

Arguments

segments

Input data from vecshift() with over_id column. Must be a data.table with required columns: cf, inizio, fine, arco, durata, over_id, and optionally additional employment attributes for detailed analysis.

analysis_level

Character string specifying analysis granularity (default: "person"):

"person": Person-level analysis with individual overlap statistics
"aggregate": Population-level aggregate overlap patterns
"detailed": Both person and aggregate analysis with comprehensive breakdowns

min_overlap_duration

Minimum duration (days) to consider as meaningful overlap (default: 1). Helps filter out very short-term overlapping contracts.

include_temporal_patterns

Logical. If TRUE (default), analyzes when overlaps occur most frequently (monthly/seasonal patterns). Requires date columns.

group_by

Optional character string specifying grouping variable (e.g., "industry", "region") for comparative overlap analysis across different categories. Must be a column name in segments (default: NULL).

temporal_aggregation

Character string specifying temporal aggregation for pattern analysis when include_temporal_patterns = TRUE (default: "month"):

"month": Monthly overlap frequency patterns
"quarter": Quarterly overlap frequency patterns
"year": Annual overlap frequency patterns

include_unemployment_context

Logical. If TRUE (default), includes context about unemployment periods for comparison with overlapping employment patterns.

Value

When analysis_level is "person", returns a list containing:

person_overlaps: data.table with person-level overlap statistics:
- cf: Person identifier
- total_employment_periods: Total employment segments (arco >= 1)
- overlapping_episodes: Count of periods with arco > 1
- single_job_periods: Count of periods with arco = 1
- overlap_frequency: Proportion of employment periods with overlaps
- avg_concurrent_jobs: Mean arco for employment periods
- max_concurrent_jobs: Maximum simultaneous jobs (peak employment complexity)
- total_overlap_days: Sum of days spent in overlapping employment
- total_employment_days: Sum of all employment days
- overlap_intensity_pct: Percentage of employment time with multiple jobs
- avg_overlap_duration: Mean duration of overlapping episodes
- max_overlap_duration: Longest overlapping episode duration
- unique_over_id_groups: Count of distinct overlapping employment episodes
- employment_complexity_score: Weighted complexity measure

When analysis_level is "aggregate", returns aggregate-level analysis.

When analysis_level is "detailed", returns comprehensive analysis including:

person_overlaps: Person-level statistics
aggregate_overlaps: Population-level aggregate patterns
overlap_duration_distribution: Distribution of overlap durations
employment_intensity_distribution: Distribution of concurrent job counts
temporal_patterns: Temporal overlap frequency (if include_temporal_patterns = TRUE)
group_comparisons: Comparative analysis by group (if group_by specified)

All results include attributes:

analysis_parameters: Analysis configuration used
overlap_validation: Data validation and quality checks specific to overlaps
computation_time: Time taken for analysis

Details

The over_id column from vecshift() identifies continuous overlapping employment periods:

over_id = 0: Unemployment periods (no active contracts)
over_id > 0: Employment periods with same value for overlapping/continuous contracts
Same over_id: All contracts belonging to same continuous overlapping time period

This function analyzes employment overlap patterns including:

Overlap Episodes: Periods where arco > 1 (multiple concurrent jobs)
Employment Intensity: How many jobs are held simultaneously
Duration Patterns: How long overlapping employment lasts
Temporal Analysis: When overlaps occur most frequently
Person-level Patterns: Individual multiple job holding behaviors
Overlap Efficiency: Time spent in multiple vs single employment

Key metrics calculated:

Total overlapping episodes: Count of periods with arco > 1
Average/maximum concurrent jobs: Employment intensity measures
Overlap duration distribution: How long multiple employment lasts
Overlap intensity percentage: Proportion of employment time with multiple jobs
Peak employment complexity: Maximum simultaneous jobs per person
Temporal patterns: Monthly/seasonal overlap frequency (optional)

Examples

if (FALSE) { # \dontrun{
library(data.table)

# Create sample employment data with overlapping periods
employment_data <- data.table(
  id = 1:12,
  cf = c(rep("PERSON001", 8), rep("PERSON002", 4)),
  INIZIO = as.Date(c("2023-01-01", "2023-02-01", "2023-02-15", "2023-05-01",
                     "2023-05-15", "2023-08-01", "2023-10-01", "2023-11-01",
                     "2023-01-15", "2023-03-01", "2023-06-01", "2023-09-01")),
  FINE = as.Date(c("2023-01-31", "2023-04-30", "2023-04-30", "2023-07-31",
                   "2023-07-15", "2023-09-30", "2023-10-31", "2023-12-31",
                   "2023-02-28", "2023-05-31", "2023-08-31", "2023-12-31")),
  prior = c(1, 0, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1),
  company = c("CompanyA", "CompanyB", "CompanyA", "CompanyC", "CompanyD",
              "CompanyE", "CompanyF", "CompanyG", "CompanyH", "CompanyI",
              "CompanyJ", "CompanyK"),
  industry = c("Tech", "Retail", "Tech", "Finance", "Tech", "Retail",
               "Finance", "Tech", "Healthcare", "Education", "Finance", "Tech")
)

# Apply vecshift to get segments with over_id
segments <- vecshift(employment_data)

# Comprehensive overlap analysis
overlap_analysis <- analyze_employment_overlaps(
  segments = segments,
  analysis_level = "detailed",
  min_overlap_duration = 1,
  include_temporal_patterns = TRUE,
  temporal_aggregation = "month"
)

# View person-level overlap patterns
print(overlap_analysis$person_overlaps)

# View aggregate overlap statistics
print(overlap_analysis$aggregate_overlaps)

# View overlap duration distribution
print(overlap_analysis$overlap_duration_distribution)

# View employment intensity patterns
print(overlap_analysis$employment_intensity_distribution)

# View temporal patterns (when overlaps occur)
if (!is.null(overlap_analysis$temporal_patterns)) {
  print(overlap_analysis$temporal_patterns)
}

# Person-level analysis only
person_overlaps <- analyze_employment_overlaps(
  segments = segments,
  analysis_level = "person",
  min_overlap_duration = 7,  # Focus on overlaps lasting at least a week
  include_temporal_patterns = FALSE
)
print(person_overlaps$person_overlaps)

# Aggregate analysis with industry comparison
industry_overlaps <- analyze_employment_overlaps(
  segments = segments,
  analysis_level = "aggregate",
  group_by = "industry",
  min_overlap_duration = 1
)
print(industry_overlaps$aggregate_overlaps)
print(industry_overlaps$group_comparisons)

# Quarterly temporal patterns
quarterly_patterns <- analyze_employment_overlaps(
  segments = segments,
  analysis_level = "detailed",
  include_temporal_patterns = TRUE,
  temporal_aggregation = "quarter"
)
print(quarterly_patterns$temporal_patterns)

# Check analysis validation and parameters
print(attr(overlap_analysis, "analysis_parameters"))
print(attr(overlap_analysis, "overlap_validation"))

# Example output interpretation:
# - person_overlaps shows which individuals have complex employment patterns
# - overlap_intensity_pct indicates how much time is spent in multiple employment
# - employment_complexity_score provides weighted measure of job juggling behavior
# - temporal_patterns reveal seasonal trends in multiple job holding
# - group_comparisons show industry-specific overlap behaviors
} # }