A creative computing approach to poetry as data

Wang, X (2018) A creative computing approach to poetry as data. PhD thesis, Bath Spa University.

Preview

Text 11560.pdf Repository Terms Apply. Download (6MB) | Preview

Abstract

With the rapid advent of emerging new services such as cloud computing, mobile technology, and social media, more and more people prefer posting their literary creations such as poems, on the Internet instead of in traditional papers. The era of Digital Humanities has truly arrived. With ever-growing concerns regarding literary data, ways to utilise and manage them has become a major concern. Many researchers have worked on that and proposed different solutions. However, owing to new challenges and creative requirements, traditional methods need adjustments. For example, most poetry data collection methods, such as surveys, are based on single target searching; that is, only relying upon keywords and themes. Thus, the result can be monotonous. Moreover, the accuracy of algorithms for poetry data analysis is no longer the only benchmark. The underlying meaning of poetry data has drawn people’s attention. Meanwhile, traditional poetry data presentation methods need to be enhanced to reflect diversity and media richness. The aim of this research is to present a Creative Computing approach to poetry data collection, analysis and presentation. The thesis demonstrates the feasibility and details the proposed methods in the following phases. Firstly, poetry data is being creatively regarded as an object with mass, volume and resistance, from an interdisciplinary perspective. A novel data relevancy rule is proposed to retrieve the closely-related data of an input, which is adapted from the Newton’s law of universal gravitation in physics. In this way, a broadened variety of data is being searched using web crawler based on multi-purpose rules. Then, the search results are filtered on the basis of buoyancy phenomenon and Ohm’s law. Secondly, with reference to chemical principles this research carries out innovative poetry data analysis based on the notion that chemical reactions always bring in brand new outcomes, despite having exactly the same elements. The mood, theme and personal reflection, after going through a piece of literature, presented difficulties for traditional data analysis. In this work, they have been investigated relying on acidity estimation, organic abstraction and oxidation-reduction reactions. Lastly, presenting the poetry analysis results through creative visualisation has been studied thanks to the elegant mathematics expressions of curves and shapes which are believed to effectively convey underlying emotions of poetry data. To illustrate this idea, a rainbow of variable spectrum and diverse types of trajectories are proposed as background and rolling titles, respectively. In summary, the proposed approach carries out manipulations on traditional poetry data processing based on models and algorithms of Creative Computing. The proposed approach was evaluated by a selected case study, where a prototype system was built for poetry analysis. Conclusions are drawn and future research is also discussed. Initial experiment results show this work contributes to an effective and Creative Computing approach to poetry data manipulation. This research has potential applications to academic research of texts, to making word recommendations for users to better comprehend literature such as poetry, a novel or drama. Furthermore, it sees the possibility of inspiring creative thinking for human art creation.

Item Type:	Thesis (PhD)
Note:	Thesis supervised by Prof Hongji Yang.
Keywords:	poetry, poetry analysis, poetry data collection, poetry data, poetry as data, poetry data analysis, creative computing, literary data, algorithms, data relevancy, buoyancy phenomenon, Ohm's law, chemical reactions, physics, Newton's second law of gravitation, mood, theme, personal reflection, creative visualisation, acidity estimation, organic abstraction, oxidation-reduction reaction, mathematics, variable spectrum, prototype system
Divisions:	Bath School of Design
Identification Number:	https://doi.org/10.17870/bathspa.00011560
Date Deposited:	16 Oct 2018 14:33
Last Modified:	21 Apr 2023 13:34
URI / Page ID:	https://researchspace.bathspa.ac.uk/id/eprint/11560

	Request a change to this item or report an issue
	Update item (repository staff only)