SNS 해시태그 분석을 통한 러닝 참여자의 디지털 경험 탐색

초록

This study aimed to explore the structural characteristics and key themes of running-related discourse on Instagram using social media big data analysis. Instagram posts were collected based on hashtags. The collected textual data were systematically preprocessed through synonym consolidation, removal of noise terms, and text normalization using regular expressions. Frequency analysis and TF–IDF analysis were first conducted to identify core keywords, followed by topic modeling based on Latent Dirichlet Allocation (LDA). The optimal number of topics was determined to be five by jointly considering semantic coherence and log perplexity measures. The results revealed five major topics: (1) smartwatch-based running coaching ecosystems, (2) running culture and event participation, (3) mobile application–based personal record management, (4) technique learning and injury prevention–oriented training, and (5) performance-oriented expert running. The findings indicate that contemporary running participation is strongly characterized by data-driven practices supported by digital devices and platforms, while also encompassing cultural and performance dimensions. This study provides empirical insights into the evolving structure of running participation by applying topic modeling to SNS big data.

키워드