~/wiki

Demographic Data Sources

Confiance : high
demographicsinseecensus-datapopulation-statisticssurvey-calibrationsynthetic-agentsinsee-detail-filesweighting-variables

Public statistical resources providing population distributions essential for calibrating synthetic agent populations and validating survey methodologies.

Key Sources

INSEE Census Detail Files

Primary Resource: Anonymized individual-level census data with weighting variables for population modeling.

2022 "Logements, individus, activité, mobilités" Dataset:

  • Individual anonymous records with demographic characteristics
  • Age, sex, housing, education, employment, mobility patterns
  • Immigration status and geographic location
  • Explicit weighting variables for population extrapolation
  • Designed for modeling and sub-population tabulations

Access: Public download from INSEE statistics portal Use Case: Foundation layer for synthetic population generation

Opinion and Attitude Surveys

ELIPSS/CDSP Panel:

  • Probabilistic French panel (~3,000 individuals in 2026)
  • Monthly questionnaires covering political opinions, environment, digital practices
  • Available through Sciences Po data portal
  • High methodological rigor for attitude modeling

European Social Survey (ESS) France:

  • Cross-national political and social attitudes
  • Free access after registration
  • Enables international benchmarking and validation

CDSP Database:

  • 400+ survey references since 1958
  • Electoral surveys and historical polling results
  • Comprehensive archive for temporal validation

Media and Market Research

ACPM Audience Data:

  • Digital brand rankings and press circulation
  • Proxy for media exposure patterns
  • Useful for understanding information consumption demographics

Implementation Strategy

Two-Layer Approach:

  1. Base Population: INSEE census files for demographic foundation
  2. Attitude Overlay: Survey sources for behavioral and opinion characteristics

Calibration Method: Use known poll results to validate synthetic populations rather than direct opinion modeling

Quality Assurance: Weighting variables and methodological transparency essential for defensible results

See also