Replication package Coming soon
Reproduce every Civica Index score from primary sources.
The Civica Index is designed to be fully reproducible. That means publishing not just the scores, but every formula, normalization step, source dataset reference, and codebook entry needed to re-derive the same numbers from scratch. The replication package ships at Beta cut-over (target: Q3 2026), once the PCA / factor analysis and confidence-interval work are finalized. The contents are described below.
Package contents
Everything needed to reproduce a Civica Index score from scratch.
- Full methodology document. An expanded version of the published methodology, including worked examples and edge-case decisions.
- Codebook. Every variable, every source, every formula — documented in a single reference table. Includes native-scale definitions and normalization bounds for each dimension.
- Processing logic. Step-by-step description of how raw source data flows into final CI scores: ingestion, normalization (fixed-bound, not observed-extremes), PCA factor weights, composite formula, confidence interval derivation, and tier classification.
- Source references. Direct links and bibliographic citations for every upstream dataset, including dataset version, release date, and coverage notes.
- Downloadable outputs. Country-level CSV covering all 197 scored jurisdictions: CI score, 90% confidence interval, rank, rank band, dimensional breakdowns, completeness flag (Full / Partial / Insufficient), and data vintage per source.
- Code (where legally permissible). The ingestion and normalization scripts from this codebase, published under an open license. Restricted upstream datasets are not redistributed — only the processing code that consumes them.
Cite this page
Cite this page
Civica. (2026). Civica Atlas Methodology — Civica Index replication package: Civica Index replication package. Civica Atlas. Retrieved May 7, 2026, from https://civicaatlas.org/civica-index/replication