Force Atlas 2 Layout

ForceAtlas2 is a continuous graph layout algorithm for handy network visualization.

This implementation is based on this paper.

Warning: Some features (especially Prevent Overlapping) are not completely implemented. I'm waiting for your pull-requests.

Example of social graph rendered with force atlas 2 layout:

Installing

Supports Python 3.3+

Install from pip:

pip install fa2l

To build and install run from source:

python setup.py install

Usage

import networkx as nx
from fa2l import force_atlas2_layout
import matplotlib.pyplot as plt

G = nx.erdos_renyi_graph(100, 0.15, directed=False)

positions = force_atlas2_layout(G,
                                iterations=1000,
                                pos_list=None,
                                node_masses=None,
                                outbound_attraction_distribution=False,
                                lin_log_mode=False,
                                prevent_overlapping=False,
                                edge_weight_influence=1.0,

                                jitter_tolerance=1.0,
                                barnes_hut_optimize=True,
                                barnes_hut_theta=0.5,

                                scaling_ratio=2.0,
                                strong_gravity_mode=False,
                                multithread=False,
                                gravity=1.0)

nx.draw_networkx(G, positions, cmap=plt.get_cmap('jet'), node_size=50, with_labels=False)
plt.show()

Features

Force Atlas 2 features these settings:

• Approximate Repulsion: Barnes Hut optimization: n² complexity to n.ln(n).
• Gravity: Attracts nodes to the center. Prevents islands from drifting away.
• Dissuade Hubs: Distributes attraction along outbound edges. Hubs attract less and thus are pushed to the borders.
• LinLog mode: Switch ForceAtlas model from lin-lin to lin-log. Makes clusters more tight.
• Prevent Overlap. WARNING! Does not work very well.
• Tolerance: How much swinging you allow. Above 1 discouraged. Lower gives less speed and more precision.
• Edge Weight Influence: How much influence you give to the edges weight. 0 is "no influence" and 1 is "normal".

Documentation

You will find all the documentation in the source code

Copyright

Copyright Eugene Bosiakov. Licensed under the GNU GPLv3.

This files are based on the java files included in Gephi (Copyright 2011 Gephi Consortium).

Also thanks to Max Shinn.