Analisis Temporal Perubahan Topik Ulasan Cyberpunk 2077 di Steam Berdasarkan Waktu Rilis Major Update Menggunakan Teknik Topic Modeling LDA, NMF dan BERTopic

Faiz, Achmad (2026) Analisis Temporal Perubahan Topik Ulasan Cyberpunk 2077 di Steam Berdasarkan Waktu Rilis Major Update Menggunakan Teknik Topic Modeling LDA, NMF dan BERTopic. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026221100-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Cyberpunk 2077 mengalami dinamika ulasan yang unik di Steam, terutama akibat rilis awal yang bermasalah dan serangkaian perbaikan besar mulai dari Patch 1.1 hingga Update 2.0. Kondisi ini menjadikannya relevan sebagai objek penelitian untuk memahami bagaimana topik yang dibahas pemain berkembang seiring perubahan kualitas game. Tugas akhir ini bertujuan untuk menganalisis karakteristik topik yang muncul serta perubahannya secara temporal di Steam berdasarkan waktu rilis major update. Metodologi yang digunakan meliputi pengumpulan data ulasan menggunakan Steam Web API, pemisahan dataset berdasarkan periode waktu rilis update yaitu PRE, P11, P15, dan U20, pra-pemrosesan teks, serta pembangunan tiga model yaitu LDA, NMF, dan BERTopic. Ketiga model dievaluasi secara kuantitatif berdasarkan skor koherensi dan skor keragaman serta interpretasi kualitatif melalui responden. Hasil evaluasi menunjukkan bahwa model LDA merupakan model yang paling optimal untuk ulasan game ini. Selanjutnya, model yang terpilih diselaraskan antarperiode waktu menggunakan cosine similiarity untuk mengetahui evolusi temporalnya serta pemberian label menggunakan pendekatan pelabelan manual, diakhir dilakukan analisis temporal didukung dengan metrik kepuasan pemain. Hasil penelitian berhasil mengidentifikasi 13 topik utama yang terdiri atas topik persisten dan transien. Analisis temporal menunjukkan bahwa pada fase awal rilis, ulasan dipenuhi topik tentang bug dan masalah teknis, susunan kata kuncinya pun menunjukkan tingginya frustasi pemain. Namun, seiring dirilisnya Patch 1.5 hingga Update 2.0, proporsi keluhan bug turun drastis, kritik terhadap ekspektasi rilis menghilang, dan tingkat kepuasan pemain pada mayoritas topik naik secara signifikan.
========================================================================================================================================
Cyberpunk 2077 has experienced unique review dynamics on Steam, primarily due to a troubled initial launch and a series of major improvements ranging from Patch 1.1 to Update 2.0. This makes it a relevant subject for studying how the topics discussed by players evolve alongside changes in the game's quality. This study aims to analyze the characteristics of emerging topics and their temporal changes on Steam, aligned with major update release dates. The methodology involved collecting review data via the Steam Web API, segmenting the dataset into time periods corresponding to updates (PRE, P11, P15, and U20), performing text preprocessing, and constructing three models, LDA, NMF, and BERTopic. These models were evaluated quantitatively based on coherence and diversity scores, as well as qualitatively through human assessment. Evaluation results indicated that LDA was the optimal model for these game reviews. Subsequently, the selected model was aligned across time periods using cosine similarity to track temporal evolution, and topics were labeled manually. Finally, a temporal analysis was conducted, supported by player satisfaction metrics. The study identified 13 key topics, comprising both persistent and transient themes. Temporal analysis revealed that during the initial release phase, reviews were dominated by topics regarding bugs and technical issues, with keywords reflecting high levels of player frustration. However, following the release of Patch 1.5 through Update 2.0, the proportion of bug-related complaints dropped drastically, criticisms regarding launch expectations vanished, and player satisfaction levels across the majority of topics rose significantly.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Analisis Temporal, Pemodelan Topik, Cyberpunk 2077, Ulasan Steam, Analisis Ulasan Game, Temporal Analysis, Topic Modeling, Cyberpunk 2077, Steam Reviews, Game Review Analysis
Subjects:	G Geography. Anthropology. Recreation > GV Recreation Leisure > GV1469.2 Computer games T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Achmad Faiz
Date Deposited:	31 Jul 2026 01:07
Last Modified:	31 Jul 2026 01:07
URI:	http://repository.its.ac.id/id/eprint/140079

Actions (login required)

View Item