Interactive force-directed graph in a jupyter notebook.
Plots an interactive force directed graph in a jupyter notebook, taking data from a dataframe of nodes and link weights.
Working with data using python in the jupyter notebook provides many options for visualising that data.
The pandas library provides convenient common visualisations, and there’s always the venerable matplotlib for more bespoke plots. The new vega-lite library is enabled by default in jupyterlab.
Sometimes we want to plot the relationships between things as a graph of linked nodes.
The networkx library can plot graphs in a notebook but the plots are static, and only really suitable for smaller data.
Being able to interactively adjust graphs really helps explore larger data.
That’s what d3fdgraph does.
Installing d3fdgraph is easy.
pip install d3fdgraph
This has been tested with Anaconda Python 3.
How To Use
To use d3fdgraph in a notebook we import the library.
d3fdgraph takes data in the form of a pandas dataframe.
That dataframe must have 3 columns:
- The first column contains the nodes at the start of a link.
- The second column contains the nodes at the end of a link.
- The third column contains a link weight number. It must be more than 0.
The larger the weight, the more strongly related the two nodes at each end of this link.
Here is an example dataframe:
Each row is a link:
- The first row is a link between apple and orange. These two are connected with a weight of 1.
- The second row is a link between apple and banana. These two are more strongly related, with a weight 2.
- The fourth row is the strongest link, with weight 3, between orange and lemon.
The column names don’t have to be source, target and weight. You can have different column names, but it is a good idea to use descriptive names.
The order of the 3 columns is what matters. d3fdgraph will use the first column as the source nodes, the second as the target, and the third as link weights.
To draw an interactive force directed graph of these nodes and links we simply pass this dataframe to the plot_force_directed_graph() function.
Here is an example of a graph made from different data.
The nodes are labelled with the names of the nodes, as described in the pandas dataframe.
The nodes are constrained from flying off the edge of the canvas.
You can use your pointer to drag nodes around to adjust the graph.
Double-clicking a node makes it jump to the centre of the canvas, from where it might depart slightly due to the simulated forces.
plot_force_directed_graph(node1_node1_weight, node_radius=3, link_distance=20, collision_scale=4)
You don’t have to use the additional parameters, but they might help improve the readability of your graph.
- node_radius is the radius of the circles drawn for each node, default 3
- link_distance is the length of each link, before divided by the link weight, default 20
- collision_scale is the multiple of the node radius that excludes other nodes, default 4
- link_width_scale scales the width of the drawn links with link weight, default is 4
A simple jupyter notebook showing how d3fdgraph can be used is here:
Note that d3fdgraph is only useful in a jupyter notebook. It isn’t useful in a terminal, and hasn’t been tested in other kinds of python notebook.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size d3fdgraph-0.34-py3-none-any.whl (5.6 kB)||File type Wheel||Python version py3||Upload date||Hashes View hashes|
|Filename, size d3fdgraph-0.34.tar.gz (5.1 kB)||File type Source||Python version None||Upload date||Hashes View hashes|