TalkBankAnno Collaborative Commentary

The goal of the TalkBankAnno project is the creation of a system to support online commentary and annotation of TalkBank materials. The project includes eight goals and we have focused so far on the first of these, which is the system for Collaborative Commentary (CC).

Collaborative Commentary

This page illustrates the use of CC for annotating transcriptions in AphasiaBank, TBIBank, and RHDBank. These tasks were used by a group of colleagues and their research staff to learn how to use CC.

This page illustrates the use of CC for annotating transcriptions in CHILDES. It was used as an assignment in a class at CMU.

The Full Project

The full TalkBankAnno project will provide support for eight types of annotation activities, each of which supports new approaches to research across wide swaths of the social sciences. Additional types may be added at a later date.

  1. Collaborative Commentary (already implemented) involves work within focused research groups who are seeking to test empirical hypotheses against language transcript data by adding commentary to select transcript segments.
  2. Transcript Coding involves adding codes across all segments in a transcript in order to compile interaction profiles.
  3. Media Coding involves adding codes or comments to data that has not yet been transcribed. It can be either a variant of Collaborative Commentary or a variant of Transcript Coding.
  4. Automated Analysis involves the use of programs that run automatically over large segments of data to create a series of new codes.
  5. Online Transcription involves the creation of a basic transcription from either audio or video media.
  6. Citizen Science opens up media and transcripts to the larger world for crowd-sourced annotation.
  7. Teaching through Commentary involves Collaborative Commentary between teachers and their students to foster learning about conversation analysis and linguistic analysis.
  8. Curation involves improvement, refinement, and correction of transcripts and codes in the database.