This package implements an open standard for working with qualitative data, as such, it has two parts: a file format/convention and this R package that facilitates working with .rock files.

The ROCK File Format

The .rock files are plain text files where a number of conventions are used to add metadata. Normally these are the following conventions:

  • The smallest 'codeable unit' is called an utterance, and utterances are separated by newline characters (i.e. every line of the file is an utterance);

  • Codes are in between double square brackets: [[code1]] and [[code2]];

  • Hierarchy in inductive code trees can be indicated using the greater than sign (>): [[parent1>child1]];

  • Utterances can have unique identifiers called 'utterance identifiers' or 'UIDs', which are unique short alphanumeric strings placed in between double square brackets after 'uid:', e.g. [[uid:73xk2q07]];

  • Deductive code trees can be specified using YAML

The rock R Package Functions

The most important functions are parse_source() to parse one source and parse_sources() to parse multiple sources simultaneously. clean_source() and clean_sources() can be used to clean sources, and prepend_ids_to_source() and prepend_ids_to_sources() can be used to quickly generate UIDs and prepend them to each utterance in a source.

For analysis, create_cooccurrence_matrix(), collapse_occurrences(), and collect_coded_fragments() can be used.