This study investigates the applicability of place attachment theory in digital environments by analyzing Steam user reviews of the narrative game <Life is Strange>. A total of 6,267 English reviews were collected, and location-related sentences were extracted and classified using BERT-based sentence embeddings. Based on established theoretical frameworks, the sentences were categorized into four components of place attachment: place dependence, affective attachment, place identity, and conative response. The analysis revealed that affective attachment (24.8%) and place identity (24.0%) were most prevalent, indicating strong emotional and identity-based engagement with in-game spaces. In contrast, place dependence and conative responses were less frequent. These findings demonstrate that narrative-driven game environments can evoke place attachment comparable to that formed in physical spaces. This study further supports the methodological potential of combining user-generated text and semantic modeling for analyzing emotional responses to digital places.

ABSTRACT
1. 서론
2. 이론적 배경
    2.1 장소애착의 개념과 구성 요소
    2.2 장소애착의 디지털 확장 가능성
    2.3 텍스트 기반 장소애착 분석의 기술적 접근
3. 연구 방법
4. 분석 결과 및 시사점
5. 결론
Acknowledgement
참고문헌

• Jung-Eun Yu(Department of Game and Interactive Media, Gachon University, Seongnam, Korea) | 유정은
• JungYoon Kim(Department of Game Media, College of IT Convergence, Gachon University, Seongnam, Korea) | 김정윤 Corresponding author