Group Concept Mapping


I’dlike to perform Group Concept Mapping (GCM) method for a research study.
This can be analysed with R (R-CMAP) run with Rstudio. But is not that transparant.

Is it possible to do the procedures necessary for Group Concept Mapping from within KNIME?

It involves:

  • occurence matrices
  • similarity matrices
  • summing all similarity matrices from each respondent into a group similarity matrix
  • MDS (on a binary square similarity matrix using euclidian distance)
  • HCA (using Ward linkage)
  • bivariate plotting

Hi @tictaco,

I am not an expert on Group Concept Mapping, but the algorithms/concepts that you have listed are available in KNIME Analytics Platform. The one task that I am not sure about of the top of my head is “summing all similarity matrices from each respondent into a group similarity matrix”. It sounds like this might involve some manual steps to achieve the summation.

Here are couple of links to get you started:

Also, do you have sample data that you could share with us to test with?



Hi Stefan,

Thank you for your answer. I’ve looked into GroupBy and Distace Matrix Calculate. I’m still unable to calculate an occurrence table as step 1 and then a similarity matrix in step 2.

I’ve attached a Demo file in xlsx format.
Worksheets ‘Statements’, ‘SortedCards’, ‘Demographics’, ‘Ratings’ and ‘RatingsScale’ are given. From ‘Occurrence Matrix’ and onward are the steps I’ve done to recreate the research proces. My goal is to automate those steps when other data (in the same first sheets) is given.

Step 1 would be to detect how many statements there are. In this example there are 80 of them. They can be detected on sheet ‘Statements’ or on sheet ‘SortedCards’ by filtering on user 1 and list every number in de data cells (A2:T14).

Step 2 is converting them to a occurrence matrix. In excel this is pretty easy.

Off course in ‘real live’ it would be convenient that KNIME is capable of counting the number of statements, and participants and iterate through those particiapants to create the similarity matrices.

Step 3: In the example every participants similarity matrix is in a sheet ( Similarity Matrix(n)’ ). When a statement is in the same pile as another statement there is a ‘1’ Else = ‘0’
In this example: participant 1 has statement 1 and 10 in the same pile. So in similarity matrix row 1 column 10 = ‘1’ and row 10 column 1 is also ‘1’.

Step 4: Then every matrix is added to the other matrices which results in sheet ‘Similarity Matrix (combined)’. highest value is ‘10’ as there are 10 participants. This can easily be checked on the diagonal because a statement is always in the same pile with itself.

DemoProject02.xlsx (1.3 MB)

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.