Density Contour (2D)

Use 2D density contour when a large x-y cloud is too dense for a plain scatter view, and read it as a density summary rather than a point-level replacement.

Figure purpose

Density contour (2D) helps summarize where points are most concentrated in a large x-y cloud. Instead of emphasizing every point equally, it emphasizes the denser parts of the relationship.

The output is a contour-style density summary of the x-y cloud: regions with more observations are represented as stronger or more prominent contour structure, while sparse regions contribute less visual emphasis than they would in a plain scatter view.

When to use or avoid

Use this figure when a plain scatter view becomes crowded enough that the overall shape is hard to read. It can be useful for seeing where observations accumulate across two numeric variables.

Avoid treating it as the best chart for all large datasets. Dense regions become easier to read, but isolated outliers and small substructures can become less visible.

For that reason, density contour often works best alongside Scatter Plot rather than as a full replacement.

Use it when your main question is "where is the cloud concentrated?" rather than "which specific points are unusual?" If the analysis depends on point labels, rare subgroups, or a small number of influential observations, keep the scatter view primary.

Required columns

  • One numeric x variable
  • One numeric y variable

The current implementation is a dedicated two-variable contour-density display, not a broad family of configurable density-estimation methods.

In the current public surface, this is a fixed summary view rather than a user-tunable density workflow. Licklider does not expose contour-bandwidth tuning, alternate kernel choices, or a menu of density estimators on this page.

This figure is still a descriptive surface. It should not be described as if it automatically solves model choice, outlier review, or subgroup interpretation.

If point-level interpretation matters, keep a Scatter Plot nearby in your reading path.

Licklider does not automatically determine from this contour view whether a sparse feature is a meaningful subgroup, whether an isolated region is an artifact, or whether the visible pattern supports a regression model. Those are follow-up interpretation tasks, not properties guaranteed by the figure itself.

Alternative figures

  • Use Scatter Plot when point-level visibility, outlier review, labels, or subgroup spotting matters more than density structure.
  • Use Regression Plot when the main question is a fitted trend and uncertainty band rather than where observations accumulate most densely.
  • Consider using both scatter and density contour when you need to balance global pattern recognition with outlier visibility.

TODO (Phase02+)

  • Expand only if the public surface later exposes clearer user-facing controls for contour tuning or density settings.