Χριστο Φόρος asked whether the difference between qualitative and quantitative information was really all that much of a problem, especially in view of mixed datasets. As I have encountered it in practice the rub is not so much between different types of data as between the two cultures of quantitative and qualitative research paradigms.
As it happens, my mix of backgrounds often found me employed consulting on statistics at the interface between quantitative and qualitative researchers. On the qual side back in the 80s and 90s we were just beginning to develop software for ethographic methods, massaging linguistic, narrative, and verbal protocols toward categorical variables and non‑parametric statistics. I worked a lot on concepts and software bridging the gap between qual and quant paradigms.
The program I spent the 80s developing and eventually submitted toward a Master’s in Psych integrated a Learning module (Slate) and a Reasoning module (Chalk). The first viewed its input stream as a two-level formal language (“words” and “phrases”) and sought to induce a grammar for the language its environment was speaking to it. The second was given propositional expressions describing universes of discourse and had to find all the conjunctions of basic qualitative features (boolean variables) satisfying those descriptions. There’s a report on this work in the following paper.