Skip to content

CLI API Reference

Entry Point

Command: panelbox Source: panelbox/cli/

Overview

PanelBox provides a command-line interface for quick model estimation and data inspection without writing Python code. The CLI supports all static panel models, GMM estimators, and multiple covariance types.

panelbox <command> [options]
Command Description
panelbox estimate Estimate a panel model from a data file
panelbox info Display data or results information

panelbox estimate

Estimate a panel model from a CSV data file.

panelbox estimate --data FILE --model TYPE --formula FORMULA --entity COL --time COL [options]

Required Arguments

Argument Description Example
--data FILE Path to CSV data file --data panel.csv
--model TYPE Model type --model fe
--formula FORMULA R-style formula --formula "invest ~ value + capital"
--entity COL Entity column name --entity firm
--time COL Time column name --time year

Optional Arguments

Argument Default Description
--output FILE, -o — Save results to file
--cov-type TYPE nonrobust Covariance type (see table below)
--format FORMAT pickle Output format: pickle or json
--verbose, -v — Print verbose output
--no-summary — Skip printing the summary table

Model Types

Value Model
pooled Pooled OLS
fe Fixed Effects
re Random Effects
between Between Estimator
fd First Difference
diff_gmm Difference GMM
sys_gmm System GMM

Covariance Types

Value Description
nonrobust Standard (default)
robust HC1 robust
hc0, hc1, hc2, hc3 Heteroskedasticity-consistent
clustered Clustered by entity
twoway Two-way clustered
driscoll_kraay Driscoll-Kraay
newey_west Newey-West HAC
pcse Panel-corrected SE

Examples

# Basic fixed effects estimation
panelbox estimate \
  --data investment.csv \
  --model fe \
  --formula "invest ~ value + capital" \
  --entity firm \
  --time year

# With robust standard errors, save results
panelbox estimate \
  --data investment.csv \
  --model fe \
  --formula "invest ~ value + capital" \
  --entity firm \
  --time year \
  --cov-type robust \
  --output results.pkl

# Random effects with clustered SE
panelbox estimate \
  --data panel.csv \
  --model re \
  --formula "y ~ x1 + x2 + x3" \
  --entity id \
  --time period \
  --cov-type clustered \
  --verbose

# Difference GMM
panelbox estimate \
  --data employment.csv \
  --model diff_gmm \
  --formula "n ~ w + k | L.n" \
  --entity id \
  --time year

panelbox info

Display information about a dataset or saved model results.

panelbox info [--data FILE | --results FILE] [options]

Arguments

Argument Description
--data FILE Path to CSV data file (mutually exclusive with --results)
--results FILE Path to saved results file (.pkl) (mutually exclusive with --data)
--entity COL Entity column (for --data mode)
--time COL Time column (for --data mode)
--verbose, -v Print verbose output

Examples

# Inspect a data file
panelbox info --data panel.csv --entity firm --time year

# Inspect saved results
panelbox info --results results.pkl --verbose

Data Info Output

When using --data, the output includes:

  • Number of observations, entities, and time periods
  • Panel balance (balanced/unbalanced)
  • Variable statistics (mean, std, min, max)
  • Missing values per column

Results Info Output

When using --results, the output includes:

  • Model type and formula
  • Coefficient estimates and standard errors
  • Fit statistics (R², AIC, BIC)
  • Diagnostic test results (if available)

Integration with Python API

The CLI is designed for quick analyses. For full control, use the Python API:

panelbox estimate \
  --data data.csv \
  --model fe \
  --formula "y ~ x1 + x2" \
  --entity id --time year \
  --cov-type robust

import pandas as pd
from panelbox.models.static import FixedEffects

data = pd.read_csv("data.csv")
model = FixedEffects(data, formula="y ~ x1 + x2",
                     entity_col="id", time_col="year")
result = model.fit(cov_type="robust")
print(result.summary())

The Python API provides access to:

  • All model types (70+)
  • Validation and diagnostic tests
  • Visualization and report generation
  • Result serialization and loading

See Also