The Role of Libraries in Preserving eSports Heritage: Archival Strategies for Competitive Gaming Records

Abstract

eSports has developed into a major socio-digital phenomenon, producing extensive streams of gameplay recordings, competitive event logs, player performance statistics, and associated multimedia artefacts. However, the long-term conservation of eSports history is constrained by several challenges, including rapid technological redundancy, intellectual property limitations, and the inherently short lifecycle of professional gaming ecosystems. To address these issues, this study proposes a digital data archival framework that investigates the function of libraries in the preservation and curation of eSports heritage, particularly within competitive gaming records. Within this framework, an eSports Archive Records Dataset comprising 15 video clip files was compiled, with source materials obtained from the Kaggle platform. Prior to analysis, the dataset underwent pre-processing using min–max normalisation to ensure consistency across metadata attributes, file formats, and indexing schemas. Feature engineering was then conducted to capture spatial configurations, temporal sequences, and motion-based patterns, thereby supporting efficient, scalable, and precise archival processes while retaining the interactive and participatory characteristics of eSports environments. For feature optimisation, the Tiki-Taka algorithm (TTA) was implemented to select the most informative attributes, enhancing analytical accuracy and model performance. In addition, deep learning architectures, specifically Convolutional-tuned Bidirectional Long Short-Term Memory networks (Con-BiLSTM), were applied for automated video analysis tasks, including metadata extraction, ingame event recognition, and highlight generation. Experimental evaluation using Python demonstrated strong system performance across multiple indicators:
retrieval accuracy (98.78%), accessibility (92.48%), preservation durability (93.57%), storage efficiency (91.72%), precision (96.37%), recall (93.98%), and F1- score (97.66%). Collectively, these results indicate that the proposed digital archival framework substantially
enhances the organisation, discoverability, and functional usability of eSports heritage data. Overall, the findings emphasise the critical role of libraries as custodians of digital cultural assets and offer a structured framework for establishing best practices in eSports archival. This approach bridges traditional preservation methodologies, community-driven documentation efforts, and contemporary AI-enabled systems for sustainable digital heritage management.