Skip to content

Quantitative scale validators accept collapsed positions for distinct values #292

Description

@rudycelekli

Problem

The bubble, slopegraph, and beeswarm validators accept a fitted scale with slope zero even when the declared values differ. That maps all values to one coordinate, so positions no longer encode the data. The corresponding axis ticks/endpoint labels can collapse too, and the public verifier still reports a clean shared scale.

Native Chromium confirms a bubble fixture with five distinct x values and ticks labelled 0 and 500 all at x=500; its y scale, area encodings, and largest-first paint order are unchanged. The public bubble CLI exits 0. Related slopegraph and beeswarm fixtures show the same root cause. This is distinct from an equal-valued dataset, which already has no derivable scale.

Proposed scope

Require a finite, nonzero fitted quantitative scale in all three callers, with normal/inverted nonzero controls and public CLI regressions for collapsed geometry. Preserve conservative degenerate-data findings, residual tolerances, and allowed perpendicular beeswarm dodge. No new metadata or chart grammar.

Codex is assisting on behalf of @rudycelekli.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions