Skip to content

The CRITT TPR-DB 3.0

version 3.0

This is the documentation site for the CRITT TPR-DB version 3.0.

The Center for Research and Innovation in Translation and Translation Technology (CRITT) is the creator and maintainer of the Translation Process Research Database (TPR-DB), which is now in its 3rd major version (3.0).

What is the TPR-DB?

The Translation Process Research Database (TPR-DB) is a tool for organizing and processing key-logging and gaze-fixation data so that it can be easily analyzed in powerful and flexible ways.

Yeah, but what is it?

The TPR-DB has three concrete resources that help you analyze your translation process data:

  • The TPR-DB web app to upload and process your data 🤖
  • Utility libraries (Python and R) to help you analyze the data 🤓
  • Documentation (you’re looking at it 😉) to help you along the way
TPR-DB 3.0 Diagram TPR-DB 3.0 Diagram
CRITT TPR-DB Flowchart

TPR-DB Workflow

The Goal

starting point = gather data

end goal = analyze data

  1. Gather key-logging data with Translog-II or Trados Qualitivity and/or gaze data with an eye tracker
  2. Upload the raw data to the TPR-DB
    1. The TPR-DB will automatically process the raw data:
      1. Segment the texts
      2. Tokenize the texts
      3. PoS tag the tokens and parse syntactic dependencies
      4. Align source and target segments
      5. (optional) Automatically align source and target words
      6. Map keystrokes to words
      7. Map fixations to words
    2. The end result of this process are SessionProps XML files
  3. Manually align words and annotate alignment groups
  4. The TPR-DB will automatically extract features to create TPR-DB data tables
  5. Use the features in the tables to visualize and analyze your data 🤓

History

The CRITT TPR-DB project started around 2010 with the data from a number of translation process studies that were recorded in Translog1 and integrated into a database under a common format. The TPR-DB was first a part of the Danish Dependency Treebanks Project at the Copenhagen Business School2 and then became a project on its own. Right from the beginning the idea was not only to provide a data repository for translation process research, but also to provide a toolkit for analyzing and visualizing the data. The TPR-DB toolkit started out with a number of functions in R to access and analyse the TPR-DB data, and subsequently developed into a browser-based toolkit with a Jupyter interface and a Python library.

The first public documentation of the CRITT TPR-DB 1.0 appeared in 2012 (Carl 2012), when Translog-II was amended to record non-European languages (Chinese, Hindi, Japanese, among others). Adding a web-interface and further toolkit functionalities resulted in the TPR-DB 2.0 some of which is documented in a Springer volume, New Directions in Empirical Translation Process Research (Carl et al. 2015). Another Springer volume appeared in 2021, Explorations in Empirical Translation Process Research (Carl 2021), when the TPR-DB was based at Kent State University.

In 2026, the TPR-DB processing chain was migrated from Perl to Python and was given a new user interface.

  1. CRITT website: Translog-II ↩

  2. The Copenhagen Dependency Treebanks ↩