IMPLEMENTATION OF TOURIST DESTINATION RECOMMENDATION SYSTEM AND EVENTS PROVISION SYSTEM FROM VARIOUS INSTAGRAM ACCOUNTS (CASE STUDY: BRAGA AND ITS VICINITY)

• Main Author: Fauzie, Fernaldi
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/75500
• Institution: Institut Teknologi Bandung

The problems studied in this final project include difficulties in determining the tourist destination to visit and getting information about events in the Braga and its vicinity which are scattered on various Instagram accounts of different tourist destinations. Ratings and reviews can be used as one of the considerations in determining the tourist destination to visit. However, not all ratings and reviews can be trusted. The main result of this final project is the implementation of a webbased application that can do filtering to obtain trustworthy ratings and reviews from Google Maps based on the length of the review and number of likes on the review using K-Means Clustering, can calculate the reputation value of tourist destinations based on the trustworthy ratings based on the length of the reviews and the number of likes on reviews, can provide dynamic tourist destination recommendations with weighted method based on the content of tourist destinations and trustworthy tourist destination ratings based on the length of the reviews and the number of likes on reviews, and can provide information about events from various Instagram accounts of different tourist destinations in the Braga and its vicinity. The system has successfully passed all functional testing, non-functional testing, reputation value calculation testing, recommendation system performance testing, and UAT. Based on the reputation value calculation testing, the system has an MAE value of 1.58. Based on the recommendation system performance testing, the system has a micro-average precision value of 0.58 and a MAP value of 0.83. UAT results show that the system meets all the user needs that have been defined previously and the system gets an overall average score of 4.29 out of 5. The proposed system has the potential to be developed further, such as adding the feature of tourist destinations recommendation based on visiting time and categories of tourist destinations and adding a more comprehensive explanation about the reputation value provided by the system.