Visualize Results of "Keygraph keyword extractor" in a network

Dead Knime-Users,

I am new in the business of text processing and after some first quickwins, I am a little bit lost with the following problem. Therefore I need your help.

I am analyzing NPS Comments and I would like to visualize a Network using the node ""Keygraph keyword extractor". The result should look like something that way:

I did some preprocessing and just left Nouns and Adjectives/Adverbs using a tag fitler node.

Am I generally right to use the Keygraph keyword extractor node? And how Do I visualize the results of the Keygraph keyword extractor node afterwards? Could someone provide an example/workflow that shows how to handle it?

Many thanks in advance



Hi Deadline,

the node Keygraph Keyword extractor can not be used for visualization. It is for keyword extraction.

To create a network between terms and visualize it you need in addition the Network Mining extension. To extract word pairs e.g. use the Term Co-occurrence node filter those pairs that co-occur frequently. Now use the nodes from the network mining extension to insert a node for each term of the frequent pairs and an edge for every pair. To insert nodes and edges you can use the Object Inserter node. Finally visualize the network by e.g. the Network Visualizer.

Cheers, Kilian


I am interested in making a graph like the one that is posed in the post. So I followed the instructions given but I think I’m doing something wrong because it generates a network with only four edges.

I had my text, I did some preprocessing to get a series of words of interest, then I calculated the co occurrence of terms and then I used Network Creator and Object Insert to establish the network but I think that’s where my error is.

The Object Insert node I configure it in the following way:
Node id column: select one of the two rows of terms previously generated by the Term Co-Occurence Counter node
Edge id column: Sentence cooccurrence

The other fields in the options tab leave them as none. The advanced options tab leaves it as the program presents it.

What I get is a network with 517 nodes and only 4 edges. When what I expect is a network with many more edges since the count of cooccurencia in the sentence that I got started from one for the different combinations of pairs of words, then I would expect that there were many more edges between nodes.

If you could explain to me, I would appreciate it.


I found my mistake



This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.