how to visualize network graph using python and pandas?

Question

I have the below dataframe that includes file number field and departements where each file number is assign to one or more departments and the number (1-0-99) represent the status of the file like

1: mean the task is done
0: the task is pending
99: the department has no access to the file

    df=pd.DataFrame({'file':[1205,2897,1171,1322,4312,2211,1242,52,443,111],
                     'finance':[1,99,0,0,1,99,1,0,1,99],
                     'IT':[1,99,0,1,99,1,0,0,99,0],
                     'marketing':[1,1,0,99,1,99,1,1,0,1]})

            file    finance      IT     marketing  
    0      1205       1          1        1     
    1      2897       99         99       1     
    2      1171       0          0        0    
    3      1322       0          1        99   
    4      4312       1          99       1     
    5      2211       99         1        99    
    6      1242       1          0        1    
    7      52         0          0        1    
    8      443        1          99       0    
    9      111        99         0        1

what i am asking is it possible to visualize the network of these files and departments using network graph??

where the expected result is :

node 1 file number 1205 is connected to the three department and all the required tasks are done
node 2 file number 2897 is connected to marketing dept
node 3 file number 1171 is connected to the three department but the tasks still pending.

Corralien · Accepted Answer · 2022-05-03 19:44:45Z

1

IIUC, you can use:

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

# df = pd.DataFrame(...)

COLORS = {0: 'red', 1: 'green'}

edges = df.melt('file', var_name='department', value_name='status').query('status != 99')
G = nx.from_pandas_edgelist(edges, source='department', target='file', edge_attr=['status'])

pos = nx.shell_layout(G, [edges['department'].unique(), edges['file'].unique()])
node_colors = ['lightcoral' if node in edges['department'].unique() else 'yellow' for node in G.nodes]
edge_colors = [COLORS[G.edges[edge]['status']] for edge in G.edges]

nx.draw_networkx(G, pos=pos, with_labels=True, node_color=node_colors, edge_color=edge_colors)
plt.show()

>>> edges
    file department  status
0   1205    finance       1
2   1171    finance       0
3   1322    finance       0
4   4312    finance       1
6   1242    finance       1
7     52    finance       0
8    443    finance       1
10  1205         IT       1
12  1171         IT       0
13  1322         IT       1
15  2211         IT       1
16  1242         IT       0
17    52         IT       0
19   111         IT       0
20  1205  marketing       1
21  2897  marketing       1
22  1171  marketing       0
24  4312  marketing       1
26  1242  marketing       1
27    52  marketing       1
28   443  marketing       0
29   111  marketing       1

answered May 3, 2022 at 19:44

Corralien

121k8 gold badges43 silver badges68 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

DevLeb2022 Over a year ago

this was exactly what i am looking for but i have several questions : can you explain the below rows:

pos = nx.shell_layout(G, [edges['department'].unique(), edges['file'].unique()]) node_colors = ['lightcoral' if node in edges['department'].unique() else 'yellow' for node in G.nodes] edge_colors = [COLORS[G.edges[edge]['status']] for edge in G.edges]

and can we change the size of department based on the number of file that he has?

Corralien Over a year ago

You have an argument node_size to change the size of a node. Read the doc

Corralien Over a year ago

pos = nx.shell_layout(...) returns the position of each nodes : inner nodes are department, outer nodes are files

Corralien Over a year ago

The rest gives a color for each node and edge.

Collectives™ on Stack Overflow

how to visualize network graph using python and pandas?

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related