Merge, subsume or group?
Andreas Johannes
I have a general question: In the Discourse Network Manager, I coded different categories or positions on postcolonial topics from newspapers, such as countries, people, etc. I then export this and visualize it with Visone. Is there a way in Visone to subsume or group positions, topics, or people again in the visualization?
For example: I import a two-mode network consisting of people and their corresponding positions. Now I’m interested in whether there could be a weighting based on the countries these people come from. So I’d like to merge the people who come from the same country. Is there any way to do that directly in Visone—by somehow combining nodes or using the grouping function? I’m really not very familiar with the technical side of the program and I’m using it for my master’s thesis, but maybe someone here can help me.
Kimmo Elo
if I understood you correctly, you could try these two options:
1) Go to "transformation" tab, then:
- level: nodes
- operation: group by
- group_by: attribute (flat)
- in "attribute", select the attribute for grouping.
Click "transform" and visone should create group nodes consisting of
nodes with the same grouping attribute value. Note, that this does not
merge nodes, just adds group nodes visible as areas.
2) Go to "transformation" tab, then:
- level: nodes
- operation: merge
- operation: contract nodes
- set the grouping attribute's "action" to "distinguish"
- check both "create multiplicity" and "collect values to list"
Now clicking "transform" merges nodes in the same group into one
node, keeping connections (a merged node's label is that of the first
in the merged set). Study the attribute manager's node/show & edit to
undertand the resulting data.
Maybe it is a good idea to make a small subset from your data and
exercise with it before moving on to work with your full data :-)
One more thing: it has been a while I have used dna to code data, so I
am not entirely sure about what it allows you to export as network
data. The examples assume your network data to provide node attributes
as well.
If something is left unclear, please do not hesitate to comment.
HTH,
Kimmo
Andreas Johannes
First: Thank you for your help
I'm sorry, but somehow this isn't really helping me, or I'm doing something wrong.
For example, in Visone I've now listed many universities as one parameter and topics as the other.
Fictional example:
-
University A is linked to the topics "Water" and "Forest"
-
University B is linked to the topics "Water" and "Law"
-
University C is linked to the topics "Forest" and "City"
That's all well and good. Now, I would like to connect the nodes of the universities based on the country they are from, for example, in order to create a network of universities from specific countries and their associated topics.
Let's say Universities A and B are both from France. As you can see from my three examples, it might turn out that French universities have an affinity for the topic "Water".
I can group University A and B (both from France), for example, simply by right-clicking and grouping them. But their links to the topic "Water" don't merge — so the strength of the connection between the group "Universities from France" and the topic "Water" isn't visualized at all.
So my question is: Is there a way to merge the links coming from grouped nodes into some sort of group-level link?
I hope you can help me with this or at least understand what I’m trying to do. By the way, the parameter "country" is not technically stored — I’m grouping the universities manually based on that.
Kimmo Elo
ok, now I understand better - or hope so :-) I think you have two
options.
Option 1: edit the data exported from dna and add the appropriate
grouping variable directly. The import this data in visone and use the
new group variable as source, the target would still be statement.
Option 2: use your original org->statement data, import it in visone
and then try the following steps:
1) Create a separate node tab (.csv) defining node grouping attributes.
Regarding your example, this table should define each university's home
country.
2) In visone, use attribute manager to import this node table (node ->
importa/export).
3) Use attribute manager to select nodes to be grouped. Select one
attribute value (e.g. France) at a time.
[Of course you can skip the steps 1-3 and select nodes to be grouped by
hand.]
4) Go to "transformation" tab (nodes still selected -> the greyed out
box "restrict to selection" should be ticked!) and set:
- operation: merge
- operation: contract nodes
- click "transform" to complete the transformation
Now you should see one node for the grouped nodes and multiple links
departing from this to one statement.
Important: now click an empty space in the graph panel visualising your
network to deselect everything.
5) In "transformation" tab, now set:
- level: links
- operation: merge
- merge: same direction [this to avoid directed links to be transformed
to undirected!]
- your dna data contains a variable for "agreement". For this
attribute, set "action" to distinguish to separate agreements and
disagreements.
- tick "create multiplicity"
- click "transform" to complete the transformation
Now you should have only one link for agreements and one for
disagreements between "countries" and "statements". You can use the new
attribute "multiplicity" to map link width to it.
Is this (close to) what you are looking for?
HTH,
Kimmo
su, 2025-03-30 kello 09:27 -0700, Andreas Johannes kirjoitti:
>
>
> First: Thank you for your help
>
> I'm sorry, but somehow this isn't really helping me, or I'm doing
> something
> wrong.
> For example, in Visone I've now listed many universities as one
> parameter
> and topics as the other.
> Fictional example:
>
> University A is linked to the topics "Water" and "Forest"
> University B is linked to the topics "Water" and "Law"