Skip to content

Visualization & Reports Tutorials

Learning Path

Prerequisites: At least one model family completed Time: 2--5 hours Level: Beginner -- Intermediate

Overview

PanelBox includes a comprehensive visualization system with 28+ chart types and an automated HTML report generator. These tutorials show you how to create publication-quality diagnostic plots, compare models visually, customize themes, and generate self-contained HTML reports that combine all your analysis in one interactive document.

The visualization system supports both Plotly (interactive) and Matplotlib (static) backends, with three built-in themes (professional, academic, presentation) and the ability to create custom themes. The report system generates standalone HTML files that can be shared with collaborators -- no Python installation needed to view them.

The existing HTML Reports Tutorial provides additional depth on the report system, including master reports and JSON export.

Visualization Notebooks

# Tutorial Level Time Colab
1 Visualization Introduction Beginner 30 min Open In Colab
2 Visual Diagnostics Beginner 45 min Open In Colab
3 Advanced Visualizations Intermediate 45 min Open In Colab
4 Automated Reports Intermediate 45 min Open In Colab

Production Notebooks

For workflows that take models from development to deployment:

# Tutorial Level Time Colab
1 Predict Fundamentals Intermediate 45 min Open In Colab
2 Save & Load Models Intermediate 45 min Open In Colab
3 Production Pipeline Advanced 60 min Open In Colab
4 Model Validation Advanced 45 min Open In Colab
5 Model Versioning Advanced 45 min Open In Colab
6 Bank LGD Case Study Advanced 60 min Open In Colab

Learning Paths

Charts (2 hours)

Essential visualization skills:

Notebooks: Visualization 1, 2

Covers chart creation, residual diagnostics, and theme selection.

Reports (3 hours)

Add automated reporting:

Notebooks: Visualization 1, 2, 4 + Production 1

Includes automated HTML report generation and prediction fundamentals.

Complete (5 hours)

Full visualization and production workflow:

Notebooks: Visualization 1--4 + Production 1--2

Adds advanced visualizations, custom themes, and model persistence.

Key Concepts Covered

  • Chart Factory: Create charts by name using ChartFactory.create()
  • Residual diagnostics: QQ plot, residual vs fitted, scale-location, leverage
  • Model comparison: Coefficient plots, forest plots, IC comparison
  • Panel charts: Entity effects, time effects, between-within decomposition
  • Themes: Professional, academic, presentation, custom themes
  • Export: PNG, SVG, PDF, HTML formats
  • HTML reports: Self-contained interactive reports
  • Master reports: Combined validation + comparison + residual reports
  • PanelExperiment: Automated multi-model analysis workflow
  • Model persistence: Save and load fitted models

Quick Example

from panelbox.visualization import create_residual_diagnostics, create_comparison_charts
from panelbox import PanelExperiment

# Residual diagnostic charts
charts = create_residual_diagnostics(results, theme='professional')
charts['qq_plot'].to_html()

# Model comparison
exp = PanelExperiment(data, formula="y ~ x1 + x2", entity_col="id", time_col="year")
exp.fit_all_models(["pooled_ols", "fixed_effects", "random_effects"])

# Master report
exp.save_master_report("master_report.html", theme='professional')

Solutions

Visualization Solutions

Tutorial Solution
01. Introduction Solution
02. Visual Diagnostics Solution
03. Advanced Visualizations Solution
04. Automated Reports Solution

Production Solutions

Tutorial Solution
01. Predict Fundamentals Solution
02. Save & Load Solution