CODEWITHSUNDEEP
networkxnetwork-analysis
Em_Rose
Em_Rose

Reputation: 25

Networkx is returning the same value for average in and out-degree for my directed Graph which is not valid

I am working with a large network. My data is stored in a relational table as an edge list which I import from a csv.file. Source and Target column values are integers. The network graph is directed. I know that the average in- degree should be higher than the average out-degree, yet networkx returns the same value for both measures. Is there something wrong with my code?

Here is my code

G= nx.from_pandas_edgelist(all, 'Source', 'Target', edge_attr=True, create_using= nx.DiGraph)

print(nx.info(G))

Dataframe = all

Source Target project_name link_name project_timestamp link_timestamp
3331 2321 A aaa 2013/01/02 2012/11/25
3332 2323 B kla 2013/01/03 2012/06/06
3332 9093 B dyr 2013/01/03 2012/02/03

Upvotes: 1

Views: 557

Answers (1)

Andrew
Andrew

Reputation: 799

There doesn't seem to be anything wrong with your code, but the average of the in-degree should be equal to the average of the out-degree in a directed graph (see handshaking lemma on Wikipedia - it mentions directed graphs in the "Definitions and Statement" section).

With the example that you've shared, for example, you have 5 nodes and three edges:

3331 -> 2321
3331 -> 2323
3332 -> 9093

From that:

Node in-degree out-degree
2321 1 0
2323 1 0
3331 0 2
3332 0 1
9093 1 0
Average 3/5 = 0.6 3/5 = 0.6

Every time an additional edge is added, the out-degree of the source node goes up by one, and the in-degree of the target goes up by one. Therefore, the sum of the in-degrees and the sum of the out-degrees remain equal to one another. Because the sums and number of nodes are always equal, the averages will be as well!

Upvotes: 2

Related Questions