ANALYSIS OF USER-GENERATED CONTENT ON TRIPADVISOR OF HOTELS IN UBUD USING TEXT MINING

• Main Author: Chiara Vau, Serena
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39132
• Institution: Institut Teknologi Bandung

User-generated content (UGCs) is becoming an important source of information for prospective consumers when making purchase decisions. Platforms are specifically developed to accommodate UGCs, with TripAdvisor, as the tourism-based platform, among them. Travelers are found to rely more on reviews by consumers than on information provided by producers for their lack of commercial intention. Attributes that are considered to be important for consumers are more likely to be mentioned in their reviews. Thus, it becomes important for management to be able to properly identify the attributes to be improved. Insights gained from UGCs can also contribute to managers’ business decisions. However, text data tends to be hard to analyze because of its large amounts and unstructured form. Text mining is a method that allows data processing to be more effective and efficient, and will be utilized in this study to collect and analyze text reviews. This study examines reviews from TripAdvisor of hotels in Ubud, Bali, using text mining approach. From the data, the objective is to identify the 10 most commonly mentioned texts (single-words and bigrams) across all hotels and calculate their correlation with those of Alaya Resort, the most reviewed hotel in Ubud, with the assumption that it represents the individual hotels. The process includes collecting data from TripAdvisor, cleaning and summarizing, analysis, and visualization of the results. Across all hotels, the 10 most frequently mentioned words are “hotel”, “staff”, “ubud”, “pool”, “stay”, “villa”, “breakfast”, “service”, “stayed”, and “beautiful”. Considering the object being reviewed is of hotels in Ubud, the word “hotel” and “ubud” is determined irrelevant for the analysis. Thus, “staff” is examined further and it is found that the word is associated positively in the reviews. The correlation test between the result across all hotels and Alaya Resort is found to be 0.914 and is statistically significant. For the bigrams, the result across all hotels are “monkey forest”, “rice fields”, “free shuttle”, “private pool”, “highly recommend”, “shuttle service”, “afternoon tea”, “pool villa”, “rice paddies”, and “3 nights”. The importance of hotels in regards to tourist attractions is implied by the result, with words such as “monkey forest” and “rice fields” as the most frequent. Correlation test that is conducted resulted in 0.705 and is statistically significant. With high correlations and statistically significant result for both single-words and bigram results, the texts mentioned across all hotels and at each hotel have high similarities in regards to important attributes mentioned by reviewers. Thus, the management of hotels in Ubud should pay attention and improve on attributes identified. Staff, being the most important aspect that should be improved on, especially in regards to their interactions with consumers. Another aspect management should pay attention to is of the hotel role in local tourism, how it is providing access to tourist attractions whether through its location or the services provided, such as shuttle service. Providing view to sceneries can also contributes in delivering tourist experience that consumers expect. It is also recommended for potential hotel businesses to find a location that is near tourist destinations and/or provides a scenery for consumers.