This is interesting. I have discovered my own variant of this in analysing some documents for a client. My initial instinct was to approach it quantitatively, but a friend suggested looking into how these things are done qualitatively in social science.

I'm still learning it and it's interesting but I can't shake the feeling that the process almost seems designed to day-dream up patterns and connections between things where there are none. On the other hand, any more quantitative approach seems like it would only be useful to confirm my pre-existing beliefs, and not significantly innovate on them.

Perhaps we need a bit of both.