emerson-escolar/mappertools

Various helper functions for doing Mapper analysis (mainly with kmapper)

mappertools

Various helper functions for doing Mapper analysis. This package has been designed mainly for use with kmapper.

Dependencies

mappertools requires:

• python
• numpy
• scipy
• scikit-learn
• pandas
• networkx
• pyclustering
• matplotlib

Install

Install from release package:

0. Navigate to the releases page
1. Download a dist package "mappertools-*.*.*.tar.gz", NOT "Source code"
2. Run pip install mappertools-*.*.*.tar.gz in command line

Alternatively, for developers install using the following.

Develop

Install from source:

git clone https://github.com/emerson-escolar/mappertools
cd mappertools
pip install -e .

Usage

1. mappertools/mapper contains several functions that may be useful for constructing mapper graphs. For example:
   • clustering.EPCover: "equalized projection cover"
   • hierarchical_clustering.HeuristicHierarchical: hierarchical clustering with automated heuristics to determine number of clusters
   • distances, filters, clustering: as labeled
2. mappertools/features contains several functions for analyzing mapper graphs
   • flare_balls: Compute "flareness" of entities in Mapper graph using the proposed definition in Escolar et al., "Mapping Firms' Locations in Technological Space"
   • flare_tree: Compute "flares" in G using the 0-persistent homology of centrality filtration.
   • mapper_stats: compute some summary statistics of a mapper graph

License

MIT