Collecting data table for documents where terms and phrases are found with JSL

kritterb

I'm using JSL to run Text Explorer in JMP Pro 13.2.1 for my text analysis.

I know I can right-click a term or phrase, select "Show Text", and get a pop-up with examples of the term within the paragraph for each document it appears in, which means that JMP Pro must somehow access or generate a list of documents for that term in that moment that it then pulls the blocks of text from to show the context. 

 

I want to use JSL to access that list and save it both with a "by" variable and without one. Preferably, this would be a table for all terms/documents so I don't end up with thousands of tables, but I'm willing to work with whatever is available. How do I go about that?

ih

Check out the document term matrix, with a little work I think it will do what you are asking for.  For example, try this (beware, you could make a huge table doing this):

 

1. In a text analysis, save the document term matrix (dtm)
2. Stack (under Table menu) the table by all of the dtm columns
3. Select and delete all rows that have a zero in the 'Data' column

Now the 'Label' column should show the individual terms and the rest of the columns are left intact, including whatever column(s) hold the documents.