Skip to content

Charts & Visualization

PanelBox includes a comprehensive visualization system with 35+ interactive chart types built on Plotly. Charts are designed for panel data workflows: residual diagnostics, model comparison, coefficient plots, time series decomposition, and more. Three built-in themes (Professional, Academic, Presentation) ensure publication-ready output.

Chart Categories

  • Residual Diagnostics

    Q-Q plots, residuals vs. fitted, scale-location, leverage plots, residual distribution, partial regression

  • Model Comparison

    Coefficient comparison, forest plots, model fit comparison, information criteria charts

  • Distribution

    Histograms, KDE plots, violin plots, box plots

  • Time Series

    Panel time series, trend lines, faceted time series

  • Panel-Specific

    Entity effects, time effects, between-within variation, panel structure

  • Econometric Tests

    ACF/PACF, unit root visualization, cointegration heatmaps, cross-sectional dependence

  • Validation

    Test overview, p-value distribution, test statistics, test comparison heatmap, validation dashboard

  • Quantile

    Quantile process plots, 3D surfaces, interactive coefficient plots

Available Charts

Residual Diagnostics

Chart Class Description
Q-Q Plot QQPlot Normality check for residuals
Residuals vs. Fitted ResidualVsFittedPlot Heteroskedasticity and nonlinearity
Scale-Location ScaleLocationPlot Variance stability
Residuals vs. Leverage ResidualVsLeveragePlot Influential observations
Residual Time Series ResidualTimeSeriesPlot Serial correlation patterns
Residual Distribution ResidualDistributionPlot Histogram + KDE of residuals
Partial Regression PartialRegressionPlot Added-variable plots

Model Comparison

Chart Class Description
Coefficient Comparison CoefficientComparisonChart Side-by-side coefficient estimates
Forest Plot ForestPlotChart Confidence interval visualization
Model Fit Comparison ModelFitComparisonChart R-squared, AIC, BIC comparison
Information Criteria InformationCriteriaChart AIC/BIC across models

Distribution

Chart Class Description
Histogram HistogramChart Frequency distribution
KDE KDEChart Kernel density estimation
Violin Plot ViolinPlotChart Distribution shape by group
Box Plot BoxPlotChart Summary statistics by group

Time Series

Chart Class Description
Panel Time Series PanelTimeSeriesChart Multi-entity time series
Trend Line TrendLineChart Trend visualization
Faceted Time Series FacetedTimeSeriesChart Small multiples by entity

Panel-Specific

Chart Class Description
Entity Effects EntityEffectsPlot Fixed/random effects by entity
Time Effects TimeEffectsPlot Time fixed effects
Between-Within BetweenWithinPlot Variance decomposition
Panel Structure PanelStructurePlot Balance and coverage

Econometric Tests

Chart Class Description
ACF/PACF ACFPACFPlot Autocorrelation diagnostics
Unit Root Test UnitRootTestPlot Stationarity visualization
Cointegration Heatmap CointegrationHeatmap Pairwise cointegration tests
Cross-Sectional Dependence CrossSectionalDependencePlot CD test visualization

Validation

Chart Class Description
Test Overview TestOverviewChart Summary of all diagnostic tests
P-Value Distribution PValueDistributionChart Distribution of test p-values
Test Statistics TestStatisticsChart Test statistic visualization
Test Comparison Heatmap TestComparisonHeatmap Cross-model test comparison
Validation Dashboard ValidationDashboard Comprehensive validation view

Quick Example

from panelbox import FixedEffects
from panelbox.visualization import ChartFactory
from panelbox.datasets import load_grunfeld

data = load_grunfeld()
model = FixedEffects("invest ~ value + capital", data, "firm", "year")
results = model.fit(cov_type="clustered")

# Create charts using ChartFactory
factory = ChartFactory(results)

# Residual diagnostics
factory.residual_vs_fitted().show()
factory.qq_plot().show()

# Entity effects
factory.entity_effects().show()

Using ChartFactory

The ChartFactory is the recommended entry point for creating charts. It automatically extracts the necessary data from model results:

from panelbox.visualization import ChartFactory

factory = ChartFactory(results)

# All chart types are available as methods
chart = factory.residual_vs_fitted()
chart = factory.coefficient_comparison([results1, results2])
chart = factory.forest_plot()
chart = factory.entity_effects()

Direct Class Usage

For more control, instantiate chart classes directly:

from panelbox.visualization import QQPlot

chart = QQPlot(residuals=results.resid)
fig = chart.create()
fig.show()

Themes

PanelBox provides three built-in themes:

Theme Best For
Professional Reports, client presentations
Academic Journal papers, dissertations
Presentation Slides, large-screen display 
from panelbox.visualization import ChartFactory, Theme

factory = ChartFactory(results, theme=Theme.ACADEMIC)
chart = factory.forest_plot()
chart.show()

Exporting Charts

# Save as HTML (interactive)
chart.write_html("chart.html")

# Save as static image
chart.write_image("chart.png", width=800, height=600)
chart.write_image("chart.pdf")
chart.write_image("chart.svg")

Detailed Guides

Tutorials

See Visualization Tutorial for interactive notebooks with Google Colab.

API Reference

See Visualization API for complete technical reference.