After splitting replies from a survey with text passages into a table with two columns (SurveyReplyID, Word), I want to be able to display and filter the results (count (distinct SurveyReplyID)) in a tree graph. Each time I select a word, all words that are also used in replies with the filtered word are shown. The user can then select a new word, and thus further reduce the realm of replies.
When the user filters 'Orange' in the first step, the tree graph shows the words 'Orange','Apple','Lemon'. A result table on the same page will show the reply SurveyReplyIDs 1, 2, 3.
If the user then selects Lemon as a second filter the result table will only show SurveyReplyID 3
The idea is to show the users words that were also mentioned with the selected words to help identify similar replys from the survey.
I have a fealing that I need to use set analysis and '=p', but I can't seem to get my head around it. Does anybody have an idea or can point me into the right direction? Thanks!