International Hellenic University
The use of social media as bibliographic citations in open access journals: a reproducible pipeline for extraction and archiving of social media based references
Abstract
dc:descriptionThe plethora of data, which keeps on growing exponentially, alongside the widespread adoption of social media platforms (SMPs), as well as the continuous expansion of scientific publishing has transformed scholarly communication. It is important to examine whether, and to what extent, academics, scholars and researchers consider SMPs as trustworthy sources in academic publications. The motivation for this thesis derived from the engagement and interaction with academic publications and student assignments and dissertations, where a shift in citation practices was observed towards non-traditional sources, including web-based and Social Media (SM) content. This study investigates the relationship between SMPs and bibliographic article citations over a five-year period (2020-2024). The primary aim of this doctoral thesis is to identify the presence of social media-based references, analyse their characteristics and examine their scope of use, accessibility and archiving status. In order to achieve this, automated pipelines were developed for downloading and renaming articles, and extracting references from them, using Python scripts generated with the aid of Artificial Intelligence (AI). The task was undertaken, with the user having no prior Python programming experience, relying partially on familiarity with HTML and CSS structures. Excluding one manually processed journal used as a descriptive baseline, 31 journals were evaluated through AI-assisted workflows. Successful downloader and renamer workflows were developed for 28 journals (90,32%), while two renaming workflows were additionally developed for three journals where automated downloading proved unsuccessful. Regarding the total time required for this feasibility case, the combined workflow execution times were generally lower than the manually recorded baseline, since the observed average operational time corresponded to approximately 15,8 seconds per file for the fully manual procedure, 13,8 seconds for the complete automated workflow corpus, and 10,8 seconds after excluding one highly time-consuming outlier case. In total, 7.666 articles from 32 Open Access (OA) journals across four disciplines (Computer Science, Education, Library Science and Physics) were analysed covering all SCImago ranking quartiles (Q1-Q4). The extracted 337.538 references were stored in a relational database and examined through SQL queries in order to detect and quantify SMP-originated citations. The findings revealed 338 instances of SM citations, corresponding to approximately one occurrence per 22,7 articles. Among the 13 SMPs that were examined, 11 were actively used. YouTube was the dominant platform, accounting for 49% of all SM citations, followed by Facebook (approximately 20%) and then Twitter (now X) with around 12,5%. Additional findings in this thesis show that 57,25% of the references remain text-based without any web locators (URLs). Digital Object Identifiers (DOIs) account for approximately 71% of all web-based references, however, they represent only around 30% of the total number of references, indicating that the majority of the dataset lacks a persistent digital identifier. The archiving analysis of SM references revealed that 6,5% of the initial posts were no longer accessible at the time of retrieval. In terms of distribution within the articles, 44% of SMP citations were in the Literature Review section, followed by the Introduction and Methodology sections (both at 17%), indicating that SM references are primarily used to provide contextual background and to support the literature, but at the same time to also support research design. The key contributions of this research include the empirical mapping of social media in scholarly references, the identification of a hybrid citation ecosystem, and the provision of cross-disciplinary and quarterly-based insights. It also introduces methodological approaches for article automated downloading and reference extraction through AI-assisted scripts. The research also highlights significant issues and deficiencies related to accessibility and archiving and emphasises the emergence of a hybrid citation ecosystem in which SM and web-based sources complement traditional scholarly material. While these sources are increasingly integrated into the research, even by high-ranked journals as this study demonstrated, the results also reveal persistent challenges related to data and metadata completeness, long-term accessibility and reproducibility. For these reasons, the study highlights the need for improved citation practices, robust archiving mechanisms, and enhanced metadata frameworks to support the sustainability and reliability of scholarly communication and therefore it proposes a metadata framework (SM Core) to support the description, the reuse and the curation of social media content in scholarly publications. Finally, this doctoral thesis provides an empirical and methodological basis for future research and development.
Degree
thesis:*- Grantor dc:publisher
- International Hellenic University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Rousidis, Dimitrios
- Ρουσίδης, Δημήτριος
Subjects
dc:subject × 24- Ανάλυση μέσων κοινωνικής δικτύωσης
- Βιβλιογραφικές παραπομπές
- Αρχειοθέτηση
- Περιοδικά ανοικτής πρόσβασης φυσικής, επιστήμης υπολογιστών, εκπαίδευσης, βιβλιοθηκονομίας
- Ροές εργασίας python με υποστήριξη τεχνητής νοημοσύνης
- Πλαίσιο μεταδεδομένων μέσων κοινωνικής δικτύωσης
- Social media analytics
- Citations
- Archiving
- Open access physics, library science, education, computer science journals
- AI-assisted python workflows
- Social media metadata framework
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Πληροφοριακά συστήματα
- ΜΜΕ και Επικοινωνίες
- Κοινωνικές Επιστήμες
- Επιστήμη πληροφόρησης
- Computer and Information Sciences
- Natural Sciences
- Information Systems
- Media and Communications
- Social Sciences
- Information science
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/62608