BeHappy

BeHappy is a small Python package that runs a collection of classic machine learning and statistics experiments with a single function call:

from BeHappy import code

code()

Calling code() runs eight experiments in order and prints (and plots) their results:

Nothing runs when you import the package. The experiments only run when you call code().

Note: code() needs an internet connection. Experiments 2, 3, 4, 6 and 8 download public datasets. Experiment 6 creates a dji/ folder and Experiment 8 creates kddcup.data_10_percent(.gz) (about 75 MB) in the current working directory. A full run takes a few minutes.

Installation

pip install behappy

The project is called behappy on PyPI, and you import it as BeHappy. pip treats the name case-insensitively, so pip install BeHappy works too.

To install from a local copy of the source:

pip install .

Usage

Python (script or terminal)

from BeHappy import code

code()

Save that as run.py and run python run.py. Each plot opens in a window, and the script continues after you close it.

Jupyter Notebook

In a cell:

%pip install behappy

Then, in a new cell:

from BeHappy import code

code()

Plots appear inline below the cell.

Google Colab

!pip install behappy

from BeHappy import code

code()

Colab may ask you to restart the runtime after installing (for example if numpy was upgraded). If it does, restart and run the second cell again.

Requirements

Python 3.9 or newer. These dependencies are installed automatically: numpy, pandas, scipy, scikit-learn, statsmodels, matplotlib and requests.

Development

pip install -e ".[dev]"
pytest

License

MIT. See LICENSE.