期刊名称：DigitCult@Scientific Journal on Digital Cultures
电子版ISSN：2531-5994
出版年度：2018
卷号：3
期号：2
页码：19-36
DOI：10.4399/97888255181532
出版社：Aracne editrice
摘要：This paper focuses on Wattpad, a social reading platform on which people can add comments in the margins of books. Analysing these comments enables the comparison between specific parts of the text and the effects they have on readers. We outline a new research programme, discussing both theoretical and practical issues in the study of Wattpad: from the identification of a methodology holding together reader response theory, cognitive literary studies, and computational text analysis, to the definition of a digital mixed method for the recognition of the linguistic and textual cues that trigger certain effects. We describe a dataset built by scraping the Wattpad website: preliminary statistics on the most commented books in the categories “Classics” and “Teen Fiction” are presented and discussed. To provide an example of the possible uses of the dataset, we introduce a simplified experiment with the sentiment analysis software Syuzhet. By comparing the “emotional arcs” produced in parallel by text and comments, we evaluate the approach and show the substantial differences between the intrinsic emotional valence of the text and the effects it produces.
其他摘要：This paper focuses on Wattpad, a social reading platform on which people can add comments in the margins of books. Analysing these comments enables the comparison between specific parts of the text and the effects they have on readers. We outline a new research programme, discussing both theoretical and practical issues in the study of Wattpad: from the identification of a methodology holding together reader response theory, cognitive literary studies, and computational text analysis, to the definition of a digital mixed method for the recognition of the linguistic and textual cues that trigger certain effects. We describe a dataset built by scraping the Wattpad website: preliminary statistics on the most commented books in the categories “Classics” and “Teen Fiction” are presented and discussed. To provide an example of the possible uses of the dataset, we introduce a simplified experiment with the sentiment analysis software Syuzhet. By comparing the “emotional arcs” produced in parallel by text and comments, we evaluate the approach and show the substantial differences between the intrinsic emotional valence of the text and the effects it produces.
