When people write about themselves over time in a diary or journal, they create a rich narrative of matters important to them. This narrative typically evolves over time and can be easily understood by humans when reading the reflective text. Understanding what people say about themselves in their reflective writing has significant potential for benefits in areas like education and mental health.
However human analysis of large quantities of reflective writing is very difficult as it is a labour intensive process. This project aims to build on the state of the art for Reflective Writing Analytics by developing computational approaches to analysing how an individual's reflective writing changes over time.
Drawing on relevant Natural Language Processing and Machine Learning technologies, a prototype for automated temporal analysis of reflective journal entries will be developed and evaluated. Implications for a specific application area (like learning or mental health) will also be explored.
- An extensive review of relevant literature will be conducted focusing on temporal aspects of reflective writing
- Leading psychological theories relating reflective writing to an application area (like learning or mental health) will be examined for their potential to unpin a computational approach
- Leading computational NLP and ML techniques will be examined for their potential to providing meaningful insights
- A pragmatic inquiry approach will be used to generate and test hypotheses linking psychological theory and computational techniques.
- The most promising approaches will be prototyped and evaluated.
- The complete process will be documented and written-up in a form appropriate to the selected project size (e.g. thesis, report, conference or journal papers, etc).
- An understanding of temporal features in reflective journals
- An understanding of most relevant psychological theories and exiting computational techniques and how they might be used together to advance the current state of the art in Reflective Writing Analytics (RWA)
- Empirical evidence of the effectiveness of hypothesised analytics techniques
- Advancing the extent to which RWA might be used within a given application area (like education or mental health)
Skills and experience
The project requires someone with solid programming ability (capable of existing software libraries to perform analysis on text) and an interest in researching the intersection between people and information. The project is not focused on development of new computational techniques, but rather the effective application of existing computational techniques to a human information interaction problem.
You may be able to apply for a research scholarship in our annual scholarship round.
Contact the supervisor for more information.