SciPyMasterPro¶
A hands-on toolkit for mastering SciPy: ten concept notebooks, an interactive Streamlit app, a tested utility library and reproducible synthetic datasets, all wired into a modern CI/CD pipeline.
Start here¶
-
Getting Started
Install with pip, uv or Docker and run your first analysis in five minutes.
-
Notebooks
Ten notebooks from descriptive statistics to curve fitting, each executed in CI.
-
Streamlit App
Twelve interactive pages covering every module.
-
API Reference
Every utility function, generated from its docstring.
What is inside¶
| Layer | Contents |
|---|---|
notebooks/, shared_notebooks/ |
10 concept notebooks + 2 SciPy-vs-statsmodels comparisons |
streamlit_app/ |
landing page + 12 interactive pages |
utils/ |
10 modules, ~4,300 lines, NumPy-style docstrings, type hints |
synthetic_data/ |
9 reproducible datasets and their generator |
tests/ |
unit, Hypothesis property, plot smoke, app, CLI and docs-example tests |
scipymasterpro/ |
the scipymasterpro command line |
Quick taste¶
import numpy as np
from scipy import stats
from utils.distribution_utils import fit_distribution, perform_ks_test
from utils.sim_utils import bootstrap_sample, compute_bootstrap_ci
data = np.random.default_rng(0).gamma(2.0, 1.5, size=400)
params = fit_distribution(data, stats.gamma)
print(perform_ks_test(data, stats.gamma, params))
boots = bootstrap_sample(data, n_iterations=2000, seed=0)
print(compute_bootstrap_ci(boots, ci=95))
Why synthetic data?¶
- concepts stay in focus, not domain noise
- every experiment is repeatable
- assumption violations can be constructed on purpose
- edge cases needed for tests are one function call away