Simba documentation

Exports and Reporting — PDF Reports, CSV Data, and Chart Images

Every model results tab in Simba supports both PDF report generation and CSV data download, giving you multiple ways to share findings with stakeholders and integrate with external tools. Additionally, Plotly charts offer built-in PNG export, and scenario/optimization results can be exported for offline analysis.

PDF Reports

Every tab on the Active Model page has an “Export to PDF” button that generates a formatted report containing chart images and data tables.

How it works:

Charts are captured as PNG images (Recharts via DOM capture, Plotly via its native export method)
A progress bar shows capture progress (e.g., “Capturing Charts for PDF: 5/10 charts”)
The report is generated client-side and downloaded automatically
Filename format: {TabName}_Report_{modelId}_{date}.pdf

Available PDF Reports

Tab	Report Contents
Media Results	Channel-level contribution charts, ROI metrics, spend/revenue breakdown
Response Curves	Saturation curves per channel with parameter summaries
Contributions	Time series decomposition showing each channel’s contribution over time
Coefficients	Posterior coefficient estimates with 94% HDI bounds
Model Statistics	R², MAPE, convergence diagnostics, goodness-of-fit metrics
Actual vs Model	Time series overlay of actual vs predicted values
Residuals	Residual plots, ACF, Q-Q diagnostics
Impulse Response	IRF charts showing variable response to shocks (VAR models)
Variance Decomposition	FEVD pie and stacked bar charts (VAR models)
Long-Run Effects	Elasticities, long-term multipliers, and path breakdown (VAR models)

CSV Data Downloads

Every tab also has a “Download CSV” button that exports the underlying data for that view.

Available CSV Exports

CSV Type	Tab / Location	Contents
`media_results`	Media Results	Per-channel metrics: contribution, spend, revenue, ROI
`contributions`	Contributions	Time series of channel contributions
`parameters`	Response Curves	Adstock and saturation parameter estimates
`curves`	Response Curves	Response curve data points (spend vs predicted response)
`coefficients`	Coefficients	Posterior coefficient estimates with credible intervals
`model_stats`	Model Statistics	R², MAPE, DIC, convergence metrics
`avm`	Actual vs Model	Actual and predicted time series values
`residuals`	Residuals	Residual values and diagnostic statistics
`irf`	Impulse Response	IRF trace data per variable (VAR models)
`fevd`	Variance Decomposition	FEVD proportions per variable and horizon (VAR models)
`long_run_effects`	Long-Run Effects	Elasticities, multipliers, path breakdown data (VAR models)
`posterior_config_grid`	Model Configuration	Full prior configuration as CSV
`optimization_results`	Budget Optimization	Optimized allocation per channel per week

Note: Response Curves downloads a ZIP file containing two CSVs (parameters.csv + curves.csv).

Chart Image Export

Charts built with Plotly.js include a built-in toolbar with a “Download plot as PNG” button. Hover over any Plotly chart to see the toolbar appear in the top-right corner.

Charts with PNG export (Plotly.js):

Response Curves (revenue curves, marginal curves)
Actual vs Model time series
Residuals (time series, ACF, Q-Q plot)
Impulse Response Function charts
Variance Decomposition charts
Portfolio Forecast (stacked area)

Charts without individual PNG export (Recharts):

Contributions (stacked area/bar)
Media Results charts
Portfolio ROI charts

Recharts charts cannot be exported individually as PNG, but they are captured and included in PDF reports.

Optimization Exports

Export	Format	Location	Contents
Optimization Results	CSV	Results header (Download icon)	Optimized spend per channel per week
Laydown Weights	CSV	Optimizer wizard → Laydown step	Weekly distribution weights
Weekly Overrides Template	CSV	Optimizer wizard → Weekly Overrides	Blank template for custom weekly allocations

See Budget Optimization for the full optimization workflow.

Scenario Exports

Export	Format	Location	Contents
Scenario Results	CSV	Scenario Planner results header	Predicted revenue per channel per period
Portfolio Scenario	JSON	Portfolio scenario results	Full state: brand templates, results, budget allocation, timestamp

The portfolio scenario JSON export includes everything needed to reproduce the scenario: all brand configurations, per-brand results, portfolio-level aggregates, and budget allocation details.

Configuration Exports

These exports help you save and reuse model configurations across projects:

Export	Format	Location	Contents
Model Configuration	JSON	Wizard → Model Details (Download Config)	Complete model setup: variables, priors, transforms, options
Lift Test Data	JSON	Wizard → Lift Test Calibration (Export)	Lift test entries with channels, values, units, cost types
Variable Selection	JSON	Wizard → Variable Selection (Export)	Variable categorizations and assignments

Configuration JSON files can be imported into a new model setup via the corresponding “Upload” or “Import” buttons.

Next Steps

Platform guides:

Measurement — Interpreting the model results you’re exporting
Budget Optimization — Optimization workflow and results
Scenario Planning — Forecasting and what-if analysis
Model Configuration — Configuration JSON details
Long-Term Effects — VAR-based exports (IRF, FEVD, LRE)

Core concepts:

Bayesian Modeling — Understanding posterior estimates in exports
Saturation Curves — Response curve data interpretation
Adstock Effects — Carryover parameters in exports

This site is open source. Improve this page.