Create a weighted graph based on Dataframe

Question

consider a data frame like this:

id	source	Target	Weight
1	A	B	1
2	A	C	2
3	A	D	3
4	A	E	4

I want to depict a graph with networkX which shows us two things: 1-Node with more connections has a larger size, respectively. 2-Edge with more weight has a thicker line in between.

Answer 1

We can set the edge_attr to Weight when we create the Graph from_pandas_edgelist then when we draw the graph we can get_edge_attributes and pass that as the width of whatever drawing operation.

For node_size we can use nx.degree to get the Degree from the Graph:

nx.degree(G)

[('A', 4), ('B', 1), ('C', 1), ('D', 1), ('E', 1)]

We can then scale up the degree by some factor since these values are going to be quite small. I've chosen a factor of 200 here, but this can be adjusted:

[d[1] * 200 for d in nx.degree(G)]

[800, 200, 200, 200, 200]

All together it can look like:

G = nx.from_pandas_edgelist(
    df,
    source='source',
    target='Target',
    edge_attr='Weight'  # Set Edge Attribute to Weight Column
)

# Get Degree values and scale
scaled_degree = [d[1] * 200 for d in nx.degree(G)]
nx.draw(G,
        # Weights Based on Column
        width=list(nx.get_edge_attributes(G, 'Weight').values()),
        # Node size based on degree
        node_size=scaled_degree,
        # Colour Based on Degree
        node_color=scaled_degree,
        # Set color map to determine colours
        cmap='rainbow',
        with_labels=True)

plt.show()

---

Setup Used:

import networkx as nx
import pandas as pd
from matplotlib import pyplot as plt

df = pd.DataFrame({
    'id': [1, 2, 3, 4],
    'source': ['A', 'A', 'A', 'A'],
    'Target': ['B', 'C', 'D', 'E'],
    'Weight': [1, 2, 3, 4]
})