I have always wondered whether it would be possible to create a methodology for analysing and categorising the data so that once this work has been done the same "recipes" can be used systematically to process the data.
Make a kind of 80/20 where 80 represents the usual repetitive cases and 20 the exceptions.
And in the exception cases, to list the exception cases by type... so that the exceptional is also treated in a similar way.
For each data item that requires non-standard or even exceptional treatment to have a questionnaire that allows classification into one or more exception categories where the best examples of resolution are demonstrated/deconstructed.
This will leave the unique cases which will have to be treated on a case by case basis.
Does the content of the book allow for this type of methodology?