Mining Goodreads. A text similarity‐based approach to measure reader absorption

  Wednesday, November 13, 2019
Mining Goodreads. A text similarity‐based approach to measure reader absorption
Verona, 13-15 November 2019
Co-working room (1st floor) in the “Polo Zanotto” of the University of Verona
Workshop organized by Simone Rebora (University of Verona, University of Basel), Piroska Lendvai (University of Basel), and Moniek Kuijpers (University of Basel).
Funded by the Swiss National Science Foundation in the “Digital Lives” programme.
Confirmed participants: Sándor Darányi (University of Borås), Christian Geng (Carstens Medizinelektronik GmbH), Oier Lopez de Lacalle (University of the Basque Country), Jean-Christophe Mensonides (École des mines d'Alès), Thomas Messerli (University of Basel), and Ute Winchenbach (University of Darmstadt).
The “Mining Goodreads” project is aimed at developing a computational approach to measure reading absorption in user-generated book reviews in English. Story world absorption is an experiential state driven by reader’s mental construction of a fictional world. Kuijpers (2014) proposed a theorization of story world absorption, distinguishing its four main compositional aspects: Attention (focused attention on the text, reducing awareness of the self and the passing of time), Transportation (the feeling of being inside the narrative), Emotional Engagement, and Mental Imagery. Kuijpers et al. (2014) developed and validated a corresponding questionnaire composed of 18 statements (Story World Absorption Scale, SWAS), which was already used in a series of studies on reader response.
In the workshop, we will present a corpus of book reviews extracted from the Goodreads platform and annotated using the SWAS categories. Based on the annotated corpus, we will train machine learning classifiers to detect different levels of absorption in millions of reviews hosted on social reading platforms.
The Programme will be structured in two introductory sessions and a presentations session (all open to the public) and an hackathon (open only to the the workshop participants)
  Wednesday 13 November Thursday 14 November Friday 15 November
10:00 – 11:30 Introduction to the Mining Goodreads project (Kuijpers, Rebora, and Lendvai) Presentations session
1. Ute Winchenbach, Introduction to the annotation tool INCEpTION
2. Sándor Darányi, Inter-annotator consistency vs. sentence embeddings: current findings
and suggestions
3. Thomas Messerli, Values on Lovelybooks
11:50 – 13:00 Extended Introduction Round (Darányi, Geng, Lopez de Lacalle, Mensonides, and Winchenbach) Hackathon
14:30 – 18:30 Hackathon Hackathon  
For information, please write to: