Methodology
Data pipeline
The atlas is built from the EJM research dataset and prepared as an analysis-ready layer for interactive browsing, filtering, and visualization.
The public-facing interface relies on a structured derivative of the research dataset that supports:
- record-level browsing
- multi-value jurisprudence and actor-type analysis
- summary statistics used across the atlas
Cleaning decisions
The current transformation makes a few deliberate simplifications so the Quarto front end can filter and chart the data reliably.
- Country values such as
US|USAare split into a display country and a preserved raw value. - Pipe-delimited multi-value fields are retained for reference and also expanded for charting.
- Status fields are condensed into a current-status field and a status-history string.
- Repeated theme and actor indicators are kept as numeric 0/1 flags for filtering.
- Numeric caretaking and representation codes are preserved, but not yet heavily featured in the UI because they still need clearer documentation.
Current limitations
Several fields in the source data deserve more explicit decoding before they become first-class interface elements.
Caretaking SystemCaretaking ActorRepresentation- The distinction between
IndigenousandIndigenous.1
Until those codes are fully documented, the atlas emphasizes better-understood fields such as legal provision type, jurisprudence framing, actor type, status, region, and year.
Next recommended improvements
- Add a documented codebook for numeric and compound categorical fields
- Produce detail pages or modal record views with fuller citations and status timelines
- Add automated tests for data-shape regressions before rendering
- Extend methodology notes with source provenance and update cadence