Feature request: node count by import

[Kimmo Elo]

Hi,

I main use visone with textual data containing e.g. term collocations. Very often same terms appear multiple times in the dataset.

Could it be possible to add a "count unique nodes"-option to the import dialog? This would add a node attribute (say: count) consisting of the frequency of each unique node in the data.

Of course I can import such a value with a node attribute table, but it would be much easier if visone would offer such a possibility when importing data.

To give a simple example, consider this link table:

--- snip ---

german -- chancellor

german -- politics

federal -- chancellor

european -- politics

chancellor -- office

--- snip ---

This would then result in a node table with node counts:

german - 2

chancellor - 3

politics - 2

federal - 1

european - 1

office - 1

Thanks in advance and many thanks for this excellent software!

Best,

Kimmo

[Müller Julian]

Hi Kimmo,

 

Isn’t this just the degree of a node? At least I can’t see any difference from your description. So you could load the (CSV) link table as a new network, compute the degree, and then export the degree again as a node attribute in the attribute manager.

 

If the problem is that the network is already loaded, there are duplicate nodes for the same value in the network and you want to compute the total degree for each value in the network, then you can use the “merge” operation for “nodes” in the “transformation” tab. If nodes having different values for a specific attribute shouldn’t be merged, you should select “distinguish” in the attribute table for these attributes. For all other attributes, you should tell visone how to determine an aggregated value for this attribute (like the mean or the sum).

If you have merged all duplicate nodes with the same value (so that only “unique” nodes remain), you can then compute the degree.

Alternatively, you can also turn this around and first compute the degree and then merge the duplicate nodes. Just make sure to sum the degree attribute during the merge operation.

 

Best,

Julian