Analisa Sentimen Menggunakan Data Twitter, Flume, Hive Pada Hadoop dan Java Untuk Deteksi Kemacetan di Jakarta

JOIN: Jurnal Online Informatika

Busman Busman¹  Nurhayati Buslim²  Tifani Shallynda Kania²  Nadika Sigit Sinatrya² 

[1] Sekolah Tinggi Ilmu Ekonomi (STIE) Gotong Royong Jakarta;Syarif Hidayatullah State Islamic University Jakarta;

关键词: Sentiment Analysis;    Big Data;    Twitter Data;    Congestion Detection;   

DOI  :  10.15575/join.v3i1.141

来源: DOAJ

【 摘 要 】

Traffic congestion big cities in Indonesia is unavoidable, especially in Jakarta. The increasing number of vehicle and the lack of public transportation is the main cause of traffic congestion in Jakarta. It disturb people activities. Government already did various efforts to resolve congestion problem, however it needs high installation, maintenance cost and need time to be implemented. Peoples often complained about traffic congestion in Jakarta by posting in Twitter which called tweets. Every tweets post are saved in API Twitter and used for sentiment analysis. It analyzed emotion of the user. Based of problems we do research how  to detect traffic congestion in Jakarta. Therefore, we try to makes Congestion Detection App. We design the app using UML diagrams. Congestion Detection App is connected with Hadoop, Flume, Hive and Derby. The app stream twitters data to colected by connecting with API Twitter. This app is Java-based application which can makes and view data tables. It  performance searching tweets data by ID and analyze traffic condition on a certain region in Jakarta. The perform sentiment analysis to a certain tweet and display the result based on the data table. The result of research is comparing Data from Congestion Detection App with data from Google Maps. We make three valus categories which consist of three colors: green for less traffic congestion have a value of 1. Orange for medium-scale traffic congestion has value of 2 and Red for heavily traffic congestion has a value of 3.  Based on three categories and value we use 4 regions for sample and comparing the values with value from Google Maps Data to get the accuracy. We got 81% average accuracy from the four samples. The result of Data from tweet sample compared with Google Maps Data. It  have big detected congestion with Congestion Detection App.

【 授权许可】

Unknown