LoginSign Up
← Back to all sparks
S

see

ANALYTICS
Velocity0.0

Visualization toolbox for statistical models in the easystats ecosystem

github.com/easystats/see ↗

see grows wherever easystats adds a diagnostic, one plot method at a time.

reasystatsdata-visualizationggplot2model-diagnostics
◆Current state
see is the visualization layer for the easystats ecosystem, supplying plot() methods for performance, parameters and datawizard objects. Each release adds methods for whatever those packages shipped — prior predictive checks, DAG diagrams, factor-analysis graphs — alongside steady theme and geom refinement.
◆Where it's heading
Growth here is downstream-driven rather than self-directed: see expands to cover new diagnostics as easystats produces them. Running alongside that is a sustained investment in presentation control — theme arguments on plot methods, elements that scale with base_size — which suits users embedding these plots in documents rather than glancing at them interactively.
◆Prediction
Expect new plot methods to keep arriving in step with performance and parameters releases, with continued theming work rather than any change in the package's scope.

◆Recent moves

  1. 1mo ago

    see 0.14.1 adds plots for prior checks and grouped means

    Three new plot methods land in one release — prior predictive checks, means by groups, and a size_arrow control for DAG diagrams — plus fixes where the type argument was ignored in check_predictions(). A textbook example of see tracking easystats feature by feature.

    View source ↗
  2. 2mo ago

    see 0.14.0 renders factor loadings as node-edge graphs

    Factor analysis and principal components gain a graph plot type rendering loadings as nodes and edges, a genuinely different way to read those models. The new halo geoms add contours around points to keep dense scatterplots legible.

    View source ↗
  3. 6mo ago

    see 0.13.0 fixes reversed plot sorting, adds theme arguments

    Daniel Lüdecke takes over as main maintainer, and several plot methods gain a theme argument. The breaking half is a fix to model_parameters() plots, where sort = "ascending" had been sorting in reverse — quietly wrong output for anyone who trusted it.

    View source ↗
  4. 11mo ago

    see 0.12.0 extends normality checks to psych factor models

    Normality checks extend to psych::fa(), psych::principal() and parameters::factor_analysis() objects, and outlier checks gain a scree plot type. The R 4.1 floor arrives via the correlation dependency rather than by choice.

    View source ↗
  5. 1y ago

    see 0.11.0 scales theme elements with base_size

    A presentation-focused release: theme functions accept pass-through arguments to ggplot2::theme(), element sizes scale properly with base_size, and theme_azurelight() joins the set. This is the theming investment that later releases keep building on.

    View source ↗
  6. 1y ago

    see 0.10.0 plots random-effect group levels for mixed models

    model_parameters() plots learn to show random-effect group levels for mixed models and gain a show_direction toggle, with better validation of size_length in p_function(). Incremental control over what makes it onto the plot.

    View source ↗