Nature uses as little as possible of anything. - Johannes Kepler
This is a Python implementation of the TDA Mapper algorithm for visualization of high-dimensional data. For complete documentation, see https://kepler-mapper.scikit-tda.org.
KeplerMapper employs approaches based on the Mapper algorithm (Singh et al.) as first described in the paper "Topological Methods for the Analysis of High Dimensional Data Sets and 3D Object Recognition".
KeplerMapper can make use of Scikit-Learn API compatible cluster and scaling algorithms.
KeplerMapper requires:
- Python (>= 3.6)
- NumPy
- Scikit-learn
Using the plotly visualizations requires a few extra libraries:
- Python-Igraph
- Plotly
- Ipywidgets
Additionally, running some of the examples requires:
- matplotlib
- umap-learn
Install KeplerMapper with pip:
pip install kmapper
To install from source:
git clone https://github.com/MLWave/kepler-mapper
cd kepler-mapper
pip install -e .
KeplerMapper adopts the scikit-learn API as much as possible, so it should feel very familiar to anyone who has used these libraries.
# Import the class
import kmapper as km
# Some sample data
from sklearn import datasets
data, labels = datasets.make_circles(n_samples=5000, noise=0.03, factor=0.3)
# Initialize
mapper = km.KeplerMapper(verbose=1)
# Fit to and transform the data
projected_data = mapper.fit_transform(data, projection=[0,1]) # X-Y axis
# Create dictionary called 'graph' with nodes, edges and meta-information
graph = mapper.map(projected_data, data, cover=km.Cover(n_cubes=10))
# Visualize it
mapper.visualize(graph, path_html="make_circles_keplermapper_output.html",
title="make_circles(n_samples=5000, noise=0.03, factor=0.3)")Standard MIT disclaimer applies, see DISCLAIMER.md for full text. Development status is Alpha.
To credit KeplerMapper in your work: https://kepler-mapper.scikit-tda.org/en/latest/#citations