TomekMatuszek / wroclaw_taxonomy

Python package for creating dendrite based on set of points using Wroclaw taxonomy method.

wroclaw_taxonomy package

About the method

This package enables user to create dendrite out of set of points based on Wroclaw taxonomy method. It uses Euclidean distance calculated from provided variables or geographic coordinates to group objects into possibly similar groups. In the following steps, each group is merged with the other closest group until the whole dataset is combined into one coherent dendrite. In the end, dendrite can be separated into several groups/clusters to classify data points based on provided variables.

Installation

You can install this package from PyPI by running this command in terminal:

pip install wroclawtaxonomy

Or download development version of the package directly from this repository:

pip install git+https://github.com/TomekMatuszek/wroclaw_taxonomy.git

Example

Basic workflow:

import wroclaw_taxonomy as wt

dendrite = wt.Dendrite(src='data/citiesPL_pop.geojson')
dendrite.calculate(columns=['lat', 'lon'], normalize=False)

dendrite.export_objects(out_file='dendrite_points.geojson')
dendrite.export_dendrite(out_file='dendrite.geojson')

plotter = wt.Plotter(dendrite)
plotter.plot()

Customizing plot:

plotter.plot(
    level=1, lines=True,
    style = {
        "markersize": 20,
        "cmap": 'gist_rainbow',
        "line_color": '#000000',
        "object_color": '#0000ff'
    }
)

Animation showing every stage of the dendrite creation:

plotter.animate(out_file='dendrite.gif', frame_duration=1)

More examples can be found in file examples.ipynb