Although it’s fair to say that most of my university coursework leaned to the theoretical side of things, I did cobble together a respectable enough background in computing, statistics, and industrial-organizational styles of systems and simulation research that most of the work I actually got paid for later on, aside from teaching, involved getting down and dirty with real world empirical data, most of it from a wide variety of bio-sciences, health sciences, and social sciences. These experiences kept practical applications to real world scientific inquiry in the forefront of my mind all through the time I developed my series of learning and reasoning programs.
As far as concrete examples go, I have a few. The more complex ones tend to come from this or that highly specialized research study, and it’s been my experience over the years that applications like that tend to bore everyone to tears but the very specialists who love that precise sort of data.
So exposition is forced to begin with simple examples, very often falling into the class of “toy worlds” problems that AI researchers of old were wont to bandy about.
You may find a series of examples like that, proceeding from the very simplest to the moderately complex, in the User Guide that I wrote up for my Theme One Program toward the end of the 1980s.
Applications of my program to Constraint Satisfaction Problems (CSPs) are briefly detailed in the following project report from the mid 1990s.