Steam平台上已累積超過六萬款的獨立遊戲，開發者仍然持續帶出更多創意吸引遊戲玩家。而遊戲玩家在決定是否購買新遊戲時，常常受到電子口碑的影響。因此，在開發遊戲時，考慮玩家偏好並精準投放有限資源，定位新遊戲開發方向是很重要的。在激烈競爭的獨立遊戲市場中，新創獨立遊戲開發者若能了解玩家偏好，找到合適的遊戲構面及元素並結合自身開發優勢，就能作為成功開發新遊戲的參考。
但新創遊戲開發者缺少市場調查的資源，因此本研究主要目的為協助新創遊戲開發者，利用Steam平台玩家評論，挖掘熱門獨立遊戲吸引玩家的因素。從擁有最多獨立遊戲數量的2D動作遊戲類別中，研究了兩款熱門的獨立遊戲《DAVE THE DIVER》及《Vampire Survivors》，並使用Python爬蟲從Steam平台汲取遊戲的評論，並將評論分為三個維度：推薦程度、遊戲時間和遊戲構面。並使用最新的BERTopic方法類聚遊戲評論，辨識玩家對於遊戲的偏好及特徵。而評估指標有：主題連貫性評估各個主題的解讀性、主題多樣性評估遊戲各構面中主題多樣化程度。
類聚主題結果判讀後發現，雖然兩款遊戲劇情、圖像、音樂及介面簡單，且皆採取重複遊玩的模式，但各自加入不同的升級元素，小幅添增玩法的複雜度，讓玩家保持新鮮感。顯示此方法能找出熱門獨立遊戲的優勢構面，且主題中的評估指標分數能判斷構面中的吸引因素重要性，提取對於新創遊戲開發者非常有幫助的資訊。

Over 60,000 independent games are launched on the Steam platform, and the novice developers are attracted by this highly competitive and expanding market. A successful indie game not only understand player preference but also make the most of their resource to make it attractive. It is well known that players are often influenced by electronic word-of-mouth when deciding whether to purchase new games. Thus, it is important to uncover player preference and focus on the proper direction using limited resources for the developers. The main objective of this study is to use the cluster analysis of the players’ review to assist novice game developers to uncover the factors that attract players. Two popular 2D action indie games are selected to illustrate the process and effectiveness of this approach. The reviews are scraped from Steam platform using Python, and segmented into three dimensions: recommendations, hours played and game aspects. BERTopic was then employed to cluster reviews, and extract various topics. The results are assessed by topic coherence and diversity, which were used to identify player preferences and characteristics for the target games. The findings of the analysis indicate that although both games put more focus on highly repetitive and simple gameplay, they each incorporate distinct upgrade elements that gradually increase difficulty and complexity, thereby maintaining player engagement. This work illustrates the important role of using text clustering for the novice game developers.

Academy of Interactive Arts & Sciences. (2023). D.I.C.E. Awards 2023 Winners. https://www.interactive.org/games/video_game_details.asp?idAward=2023&idGame=1871

Alamsyah, A., & Girawan, N. D. (2023). Improving Clothing Product Quality and Reducing Waste Based on Consumer Review Using RoBERTa and BERTopic Language Model. Big Data and Cognitive Computing, 7(4), 168.

An, Y., Oh, H., & Lee, J. (2023). Marketing Insights from Reviews Using Topic Modeling with BERTopic and Deep Clustering Network. Applied Sciences, 13(16), 9443.

Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent Dirichlet allocation [Article]. Journal of Machine Learning Research, 3(4-5), 993-1022. https://www.scopus.com/inward/record.uri?eid=2-s2.0-0141607824&partnerID=40&md5=505ce8839ae28d1cb56a7ff91bd0ad2d

British Academy Games Awards. (2023). 2023 BAFTA Games Awards: The Winners. https://www.bafta.org/games/awards/2023-nominations-winners

British Academy Games Awards. (2024). 20th BAFTA Games Awards: The Winners & Nominations. https://www.bafta.org/games/awards/2024-nominations-winners

Brunt, C. S., King, A. S., & King, J. T. (2020). The influence of user-generated content on video game demand. Journal of Cultural Economics, 44(1), 35-56.

Cai, S., Zhang, S., Zhu, L., & Jia, Y. (2023). A Data Analysis of Video Game Reviews on Steam. In X.-S. Yang, R. S. Sherratt, N. Dey, & A. Joshi, Proceedings of Eighth International Congress on Information and Communication Technology Singapore.

Chan, L. M., & Salaba, A. (2015). Cataloging and classification: an introduction. Rowman & Littlefield.

Church, K. W. (2017). Word2Vec. Natural Language Engineering, 23(1), 155-162.

Clarke, R., Lee, J., & Clark, N. (2015). Why Video Game Genres Fail: A Classificatory Analysis. Games and Culture, 12.

Dieng, A. B., Ruiz, F. J. R., & Blei, D. M. (2019). Topic Modeling in Embedding Spaces. arXiv:1907.04907. Retrieved July 01, 2019, from https://ui.adsabs.harvard.edu/abs/2019arXiv190704907D

Egger, R., & Yu, J. (2022). A Topic Modeling Comparison Between LDA, NMF, Top2Vec, and BERTopic to Demystify Twitter Posts [Methods]. Frontiers in Sociology, 7.

Grootendorst, M. (2022). BERTopic: Neural topic modeling with a class-based TF-IDF procedure. arXiv e-prints, arXiv:2203.05794.

Gu, X., & Kim, S. (2015, 9-13 Nov. 2015). "What Parts of Your Apps are Loved by Users?" (T). 2015 30th IEEE/ACM International Conference on Automated Software Engineering (ASE),

Guan, R., Zhang, H., Liang, Y., Giunchiglia, F., Huang, L., & Feng, X. (2022). Deep Feature-Based Text Clustering and its Explanation. IEEE Transactions on Knowledge and Data Engineering, 34(8), 3669-3680.

Guzsvinecz, T. (2022). The correlation between positive reviews, playtime, design and game mechanics in souls-like role-playing video games. Multimedia Tools and Applications, 82(3), 4641-4670.

Heo, S., & Park, J. (2021). Are you satisfied or satiated by the games you play? An empirical study about game play and purchase patterns by genres. Telematics and Informatics, 59, 101550.

Hichibe, N. (2009). Possibilities of Sustainable Small-Scale Game Development. Journal of Digital Games Research, 3(2), 171-183.

Jacob, D., & Ming-Wei, C. (2018). Open Sourcing BERT: State-of-the-Art Pre-training for Natural Language Processing. https://blog.research.google/2018/11/open-sourcing-bert-state-of-art-pre.html

Jeffrey, P., Richard, S., & Christopher, D. M. (2014). GloVe: Global Vectors for Word Representation. https://nlp.stanford.edu/projects/glove/

Jurafsky, D., & Martin, J. H. (2023). Speech and Language Processing. https://dudeperf3ct.github.io/lstm/gru/nlp/2019/01/28/Force-of-LSTM-and-GRU/

Kamal, A. S. b. A., Saaidin, S., & Kassim, M. (2020, 20-20 June 2020). Recommender System: Rating predictions of Steam Games Based on Genre and Topic Modelling. 2020 IEEE International Conference on Automatic Control and Intelligent Systems (I2CACIS),

Kan, M. (2023). Steam Reaches 10 Million Concurrent Active Players for the First Time. PCMag. https://www.pcmag.com/news/steam-reaches-10-million-concurrent-active-players-for-the-first-time

Kim, D., Seo, D., Cho, S., & Kang, P. (2019). Multi-co-training for document classification using various document representations: TF-IDF, LDA, and Doc2Vec. Information Sciences, 477, 15-29.

Klepek, P. (2022). How ‘Vampire Survivors’ Went From Obscurity to 27,000 People Playing at Once. https://www.vice.com/en/article/z3nyze/how-vampire-survivors-went-from-obscurity-to-27000-people-playing-at-once

Kwak, M., Park, J. S., & Shon, J. G. (2022). Identifying Critical Factors for Successful Games by Applying Topic Modeling. JOURNAL OF INFORMATION PROCESSING SYSTEMS, 18(1), 130-145.

Laurence, Hermawan, A., Bernarto, I., & Antonio, F. (2023). Video Game Engagement: A Passkey to the Intentions of Continue Playing, Purchasing Virtual Items, and Player Recruitment (3Ps). International Journal of Computer Games Technology, 2023, 2648097.

Li, X. Z., Zhang, Z. Y., & Stefanidis, K. (2021). A Data-Driven Approach for Video Game Playability Analysis Based on Players' Reviews. Information, 12(3), 129.

Lipkin, N. D. (2012). Examining Indie's Independence: The Meaning of "Indie" Games, the Politics of Production, and Mainstream Cooptation.

Loveridge, S. (2022). Here are all the Golden Joystick Awards 2022 winners. https://www.gamesradar.com/golden-joystick-awards-2022-winners/

Mahmoud, H. A. H., Hafez, A. M., & Alabdulkreem, E. (2023). Language-Independent Text Tokenization Using Unsupervised Deep Learning. Intelligent Automation & Soft Computing, 35(1).

McInnes, L., Healy, J., & Astels, S. (2017). hdbscan: Hierarchical density based clustering. The Journal of Open Source Software, 2.

McInnes, L., Healy, J., & Melville, J. (2018). UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. arXiv:1802.03426. Retrieved February 01, 2018

Mendonça, M., & Figueira, Á. (2024). Topic Extraction: BERTopic’s Insight into the 117th Congress’s Twitterverse. Informatics, 11(1), 8.

Ngan, L. (2023). Indie RPG and sushi restaurant sim Dave the Diver reaches 1m sales. https://www.eurogamer.net/indie-rpg-and-sushi-restaurant-sim-dave-the-diver-reaches-1m-sales

NLTK Team. (2023). Natural Language Toolkit. https://www.nltk.org/

Oh, Y. K., Yi, J., & Kim, J. (2023). What enhances or worsens the user-generated metaverse experience? An application of BERTopic to Roblox user eWOM. Internet Research, ahead-of-print(ahead-of-print).

Organisciak, P., & Twidale, M. (2015). When the elevator pitch meets the subject heading: How mixtures of other documents can describe what a document is about. Proceedings of the American Society for Information Science and Technology, 51(1), 1-9.

Poncle. (2021). Vampire Survivors Offical Site. https://poncle.games/vampire-survivors

Rani, R., & Lobiyal, D. K. (2022). Performance evaluation of text-mining models with Hindi stopwords lists. Journal of King Saud University - Computer and Information Sciences, 34(6, Part A), 2771-2786.

Reimers, N., & Gurevych, I. (2019). Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. arXiv:1908.10084. Retrieved August 01, 2019, from

Rey-Moreno, M., Sánchez-Franco, M. J., & Rey-Tienda, M. D. (2023). Examining transaction-specific satisfaction and trust in Airbnb and hotels. An application of BERTopic and Zero-shot text classification [Article]. Tourism & Management Studies, 19(2), 21-37.

Röder, M., Both, A., & Hinneburg, A. (2015). Exploring the Space of Topic Coherence Measures Proceedings of the Eighth ACM International Conference on Web Search and Data Mining, Shanghai, China.

Sheu, J.-J., & Chu, K.-T. (2017). Mining association rules between positive word-of-mouth on social network sites and consumer acceptance: A study for derivative product of animations, comics, and games. Telematics and Informatics, 34(4), 22-33.

Sheu, J.-J., Chu, K.-T., & Wang, S.-M. (2017). The associate impact of individual internal experiences and reference groups on buying behavior: A case study of animations, comics, and games consumers. Telematics and Informatics, 34(4), 314-325.

Simon, B. (2012). Indie Eh? Some Kind of Game Studies. Indie, Eh?, 7(2013).

Slxu.public. (2009). Smoothed LDA.png. https://upload.wikimedia.org/wikipedia/commons/4/4d/Smoothed_LDA.png

Steam. Introducing Steam Reviews. https://store.steampowered.com/reviews/

Steam. Introducing Steam Tags, A Powerful New Way to Shop For Games. https://store.steampowered.com/tag/

Steam. (2023a). The 2023 Steam Awards. https://store.steampowered.com/steamawards/?snr=1_5_9__12

Steam. (2023b). Indie Tagged Games. https://store.steampowered.com/tags/en/Indie/

Tang, W., Bu, H., Zuo, Y., & Wu, J. (2024). Unlocking the power of the topic content in news headlines: BERTopic for predicting Chinese corporate bond defaults. Finance Research Letters, 62, 105062.

Tang, Z., Pan, X., & Gu, Z. (2024). Analyzing public demands on China’s online government inquiry platform: A BERTopic-Based topic modeling study. PLOS ONE, 19(2), e0296855.

Thalemann, R., Wölfling, K., & Grüsser, S. M. (2007). Specific cue reactivity on computer game-related cues in excessive gamers. BEHAVIORAL NEUROSCIENCE, 121(3), 614-618.

The New York Videogame Critics Circle. (2023). 12th Annual New York Game Awards Winners (2023). https://nygamecritics.com/new-york-game-awards-archives/

Tong, X. E. (2021). Positioning game review as a crucial element of game user feedback in the ongoing development of independent video games. Computers in Human Behavior Reports, 3, 100077.

Zhang, Z., Fang, M., Chen, L., & Namazi Rad, M. R. (2022, July). Is Neural Topic Modelling Better than Clustering? An Empirical Study on Clustering with Contextual Embeddings for Topics. In M. Carpuat, M.-C. de Marneffe, & I. V. Meza Ruiz, Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies Seattle, United States.

林沛澧. (2021). 臺灣獨立遊戲發展脈絡之初探研究. 靜宜人文社會學報, 15.

翁榛憶. (2019). 電玩遊戲與在地歷史的交會：以《返校Detention》為例. 文化研究季刊(165), 3-20.

新頭殼. (2021). 叫好又叫座！創夢領投之國產遊戲《廖添丁-稀代兇賊の最期》上架首日登Steam銷售第二. https://newtalk.tw/news/view/2021-11-05/662070

資誠聯合會計師事務所. (2023). 2023‑2027臺灣娛樂暨媒體業展望. https://www.pwc.tw/zh/publications/topic-report/taiwan-enm-outlook.html

關鍵評論網. (2021). 【專訪】赤燭遊戲工作室：從《返校》到《還願》，「台式恐怖」如何引起跨國共鳴？. https://www.thenewslens.com/article/150327