Edge Betweenness

[pc524]

Hello,

I was trying to calculate Edge Betweenness, and saw this needs to be done through a plug-in. However, it seems that the existing plug-in is no longer compatible with the newer versions of Gephi -- is there any chance to use the plug-in with the latest versions of Gephi, or even having Edge Betweenness implemented in the main package? That would be really helpful! (I am sure Egde Betweenness could be beneficial to many users...).

Many thanks
Paolo

[xcfcode]

Have you solved this problem?

[pc524]

Hello,

No, I haven't. The Edge Betweenness plug-in doesn't seem to be compatible with the latest version of Gephi still. Does anyone know what it would take to make it compatible? It is probably not so a big task, but unfortunately I am not a programmer... (I wish I was!)

[mbastian]

Hi, indeed this plugin was never migrated to 0.9+ versions. I took the time to update it but I would like some assistance to check the correctness of the implementation. Making sure it adheres to the scientific definitions, namely

• Correct values for both directed and undirected graphs
• Correct way of normalizing

Would someone be able to try this out and report back? Much appreciated!

Here is a development version you can install in Gephi via Plugins > Downloaded after unzipping
edge-betweenness-metric-1.1.0.nbm.zip

[pc524]

Many thanks, Mathieu, for taking the time to update this plugin. My collaborators and I will run some checks and report back. Paolo

…

________________________________ From: Mathieu Bastian ***@***.***> Sent: 29 June 2022 09:27 To: gephi/gephi-plugins ***@***.***> Cc: Paolo Campana ***@***.***>; Author ***@***.***> Subject: Re: [gephi/gephi-plugins] Edge Betweenness (#236) Hi, indeed this plugin was never migrated to 0.9+ versions. I took the time to update it but I would like some assistance to check the correctness of the implementation. Making sure it adheres to the scientific definitions, namely * Correct values for both directed and undirected graphs * Correct way of normalizing Would someone be able to try this out and report back? Much appreciated! — Reply to this email directly, view it on GitHub<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fgephi%2Fgephi-plugins%2Fissues%2F236%23issuecomment-1169684679&data=05%7C01%7Cpc524%40universityofcambridgecloud.onmicrosoft.com%7Cb4ca56175a8c4ce9f23708da59a92ea6%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637920880409535773%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=FDSJBMd0X8TZxzQ19T7wc25n9JvXzPUPbV40%2Fgs863w%3D&reserved=0>, or unsubscribe<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAOQGZGNOKXIYOZRFMEJRTPLVRQCGLANCNFSM5GYCZKCQ&data=05%7C01%7Cpc524%40universityofcambridgecloud.onmicrosoft.com%7Cb4ca56175a8c4ce9f23708da59a92ea6%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637920880409535773%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=d0Mf%2B9s9ujEefeMpgrWS8jWoYuPdkOxSNeEh%2Bc%2BaFzk%3D&reserved=0>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[mbastian]

@pc524 Hi there, any update from your side? Thx

[pc524]

Hi Mathieu, Andrea Giovannetti, an econometrician and a collaborator of mine, has checked the working of the plug-in. The undirected edge betweenness works perfectly fine. Andrea will now run some checks on the directed version. So far so good! Best wishes Paolo

…

________________________________ From: Mathieu Bastian ***@***.***> Sent: 02 September 2022 07:22 To: gephi/gephi-plugins ***@***.***> Cc: Paolo Campana ***@***.***>; Mention ***@***.***> Subject: Re: [gephi/gephi-plugins] Edge Betweenness (#236) @pc524<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fpc524&data=05%7C01%7Cpc524%40universityofcambridgecloud.onmicrosoft.com%7Cd62d6f1f224a4357c94708da8cab84a4%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637976965532893413%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=bOYz%2FthIkPOxPtxT4X8HADxLDUOPLvwcwlTGAg3GfYE%3D&reserved=0> Hi there, any update from your side? Thx — Reply to this email directly, view it on GitHub<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fgephi%2Fgephi-plugins%2Fissues%2F236%23issuecomment-1235113814&data=05%7C01%7Cpc524%40universityofcambridgecloud.onmicrosoft.com%7Cd62d6f1f224a4357c94708da8cab84a4%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637976965532893413%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=kNSIG76hW71ch5x%2BUN%2BkcA417DE%2Bu69eZpsHYv0MsGc%3D&reserved=0>, or unsubscribe<https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAOQGZGPYEHMDW7TV6D7NFWDV4GMKLANCNFSM5GYCZKCQ&data=05%7C01%7Cpc524%40universityofcambridgecloud.onmicrosoft.com%7Cd62d6f1f224a4357c94708da8cab84a4%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637976965532893413%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=9o6kS9JWLaWPFazOdRM%2FGqIjceRxMCGcMwR%2FV21RqhA%3D&reserved=0>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[mbastian]

Hi @pc524, any updates? Thanks!

[giovannetti87]

Hi Mathieu,

Hope all is well. Thanks for taking our query so promptly, we warmly appreciate! I am afraid I am unable to match the vector of edge betweenness centralities generated with the porting you kindly provided us with an in-built function of R studio IGRAPH. I tried the script on a directed graph with multiple repeating edges. In gephi, as merging strategy, I asked not to merge the repeated links. Similarly, in Igraph, I used the following import commands:

import data.csv (enclosed) as data
net_test = as.network(data[c('Source', 'Target')], multiple=TRUE, directed=TRUE)
edge_list_dataframe = as.data.frame(edge_list_full)
edge_list_dataframe['edge_betweenness'] = edge_betweenness(
graph.data.frame(net_test),
directed = TRUE,
weights = NULL,
cutoff = -1
)

an ordered/standardized scatterplot of the centralities computed with the two measures gives the following;

I am likely getting something wrong (I will investigate how IGRAPH tuckles multiplicity of linkages), but I am unsure on what I should work on

thanks again,

please let us know if we can help further

data.csv

[mbastian]

Hi @giovannetti87 and thanks for the report. I'm curious if you could isolate the issue a bit further?

• Do you see mismatch even for graphs that don't have any parallel edges?
• What about normalization? Do you see a match different when turned on/off?

[giovannetti87]

Hi @mbastian, I'll check and report back

[giovannetti87]

Hi @mbastian . I had a thought about the issue and run another few tests to compare results between Igraph and gephi.
I can confirm that with a simple network made of 2 separated communities connected by a node everything works fine

S| T
1 | 2
3 | 2
4 | 2
2 | 5
5 | 6
6 | 7
6 | 8
6 | 9

As said, the structure above is two islands connected by a bridge. The problems start when we add a second bridge overarching between the two communities, say, a link between 9 and 2

S| T
1 | 2
3 | 2
4 | 2
2 | 5
5 | 6
6 | 7
6 | 8
6 | 9
9 | 2

The centrality in this case is as follows, with R (I studio) and G ( Gephi)

R | G
8 8.5
8 8.5
8 8.5
10.5 13.0
9.5 12.0
8.0 8.5
8.0 8.5
9.5 11.5
10.5 12.5

I think standardization wouldn't work because link [2,5] and [9,2] give inconsistent results between the two apps. Interestingly, these two links are part of the two alternative bridges.

Hope it helps!