Installation¶

This page covers how to install MCPower, its optional dependency groups, and how to verify that everything is working.

Basic Installation¶

pip install mcpower

This installs the core package with all required dependencies. It is sufficient for standard power analyses using linear regression models with continuous, binary, and factor variables.

Updating¶

pip install --upgrade mcpower

To check your current version:

pip show mcpower

Optional Dependency Groups¶

MCPower provides optional dependency groups for features that require additional packages:

Group	Command	What It Adds
Core (no extras)	`pip install mcpower`	Standard power analysis (OLS, factors, scenarios, mixed models)
`[pandas]`	`pip install mcpower[pandas]`	pandas DataFrame support for `upload_data()`
`[lme]`	`pip install mcpower[lme]`	statsmodels (legacy LME fallback)
`[all]`	`pip install mcpower[all]`	pandas (DataFrame support)

All Optional Dependencies¶

pip install mcpower[all]

Installs pandas for DataFrame support.

Requirements¶

Python: >= 3.10

Core dependencies (installed automatically):

Package	Purpose
NumPy	Array operations, random number generation
matplotlib	Power curve plots and visualizations
joblib	Parallel execution for mixed-effects models
tqdm	Progress bars for simulation runs

Optional dependencies:

Package	Install With	Purpose
pandas	`mcpower[pandas]`	DataFrame input for `upload_data()`
statsmodels	`mcpower[lme]`	Legacy LME solver fallback

C++ Backend¶

MCPower includes a native C++ backend built with pybind11 and Eigen for high-performance computation. During installation, pip install mcpower automatically compiles this backend using scikit-build-core and CMake.

The C++ backend is required – it provides ~200x speedups for mixed-effects models and handles all distribution functions via Boost.Math. If compilation fails (e.g., no C++ compiler available), MCPower will not work.

For more details, see Performance & Backends.

Verifying Your Installation¶

Run a quick power analysis to confirm everything is working:

from mcpower import MCPower

model = MCPower("y = x1 + x2")
model.set_simulations(400)
model.set_effects("x1=0.5, x2=0.3")
result = model.find_power(sample_size=100, return_results=True, progress_callback=False)

print(f"MCPower is working. Power for x1: {result['results']['individual_powers']['x1']:.1f}%")

Effects: x1=0.5, x2=0.3
Model settings applied successfully

================================================================================
MONTE CARLO POWER ANALYSIS RESULTS
================================================================================

Power Analysis Results (N=100):
Test                                     Power    Target   Status  
-------------------------------------------------------------------
overall                                  100.0    80       ✓       
x1                                       99.5     80       ✓       
x2                                       80.5     80       ✓       

Result: 3/3 tests achieved target power
MCPower is working. Power for x1: 99.5%

Desktop GUI Application¶

If you prefer a graphical interface, MCPower GUI is a standalone desktop application for Windows, Linux, and macOS. Download ready-to-run executables from the releases page. No Python installation required.