Wanted: text analysis capability

I think this has been suggested before, perhaps by me, but here goes again :bulb: .

I would love to do text analysis in Scrivener (like NVivo for those of you who have used it). It seems to me that the tools needed are already there, and it fits with the philosophy behind Scrivener (so KB, please be gentle on me :smiley: ).

Fundamentally, I would like to highlight text, add comments, and tie that to a keyword (known in NVivo as codes and nodes). That’s pretty much it, but there are some features that would have to accompany this:

  • It should be easier to organise keywords into trees, multiple trees maybe;
  • I need to be able to turn these highlights on and off --as you can imagine, it can get pretty messy in there;
  • I need to select a keyword and get a list of the documents, and highlighted passages “pointed” at by those keywords;
  • And for a killer blow, I would like to link keywords to Scapple so that I can draw a semantic web.

Like I said, this is just using the stuff that is already there. No point letting KB get bored, is there?

Without wishing to pre-empt anything KB may have to say on the subject, I have/have had a few MA students — also wanting to move on to PhD — who have needed text analysis. I have pointed them in the direction of TAMS Analyser, which, from what I have read, is equally capable and less frustrating than Nvivo, and has the huge advantage that it is free.

Furthermore, once you have decided on your tagging scheme for TAMS, you can tag a text in Scrivener, as the tags — like Scriv in-line annotations when exported — consist of bracketed elements, and TAMS imports .txt and .rtf files and does the rest of the analysis and graphs the results as required. Using a keyboard macro app like Typeit4Me, Typinator or TextExpander means you could easily set up shortcuts for each of your opening and closing tags.

I would also guess, that it would be possible, using labelled inline annotations in Scrivener combined with a suitable replacement setting on Compile, to use the existing annotation system to do your tagging work.

Mark

Just think about it. TAMS forces you to write your tags in the text. In other words it destroys the readability of it. They are very, very intrusive. And if you want a hierarchy of codes, well, you have to add the whole hierarchy.

TAMS may work for the analysis of text where readability doesn’t figure, but it is no good for analysing your own writing. Writers of case studies would appreciate this. Scrivener, on the other hand, could be just the ticket.

The surface stuff may already be there, but I shudder to think what might have to be done under the surface to add the functionality you describe. I’m no coder, but I’ve come to realise that things that look easy are often very far from that in reality!

I’ve done some text coding in my time, and I’ve used TAMS a little. It does get fairly unwieldy with a lot of text. I turned to pencil and paper …

Cheers, Martin.