WP 3: Text mining and citizen science

The main aim of WP3 is to use co-design methods to investigate the experiences of cancer patients and their significant others, drawing upon novel text mining and qualitative analysis methodologies, to improve outcomes for cancer patients and their families and enhance the quality of a conversation tool to be used by citizens, cancer patients and clinicians across Europe.

work package 3 tasks


Data collection and cleaning of corpora

This will be sourced from existing Patient Experience Monitor (PEM) available in multiple countries including the UK and Netherlands and other corpora available to the partners who have prior expertise and contacts to source such data.


Developing natural language processing (NLP) and machine learning (ML) software prototypes

These will automatically annotate the corpus data, initially with an existing off the shelf NLP toolset, and subsequent iterations will incorporate techniques from tasks 3 and 4


Develop novel entity linking techniques to link citizen’s life-world vocabulary to technical professional vocabulary

This task will be carried out in conjunction with the PPI Board who will contribute to defining key terms which will inform the development of the vocabulary.


Extend aspect-based sentiment analysis techniques for the cancer narratives domain

Service-user support will guide process to ensure its relevance to the three target cancer groups.


Manual annotation and analysis of metaphor in representative subsets of the narratives

This task will involve the collaboration of service-users to establish cultural sensitivity.


Create new versions of the conversation tool ‘Metaphor Menu’

Versions will be in English, Danish, Dutch and Spanish, based on a selection of metaphors from Task 5 that members of the PPI Board will regard as most helpful and/or representative.

Paul Rayson


Work package leader

Interested to learn more about the project?