Issue Date
2023
Abstract
The study analyzes the travel patterns of X/Twitter users in the Philippines by developing a probabilistic model to efficiently classify publicly posted tweets with geo-location that were retrieved from the X/Twitter Streaming API. Using probabilistic topic models specifically the latent Dirichlet allocation model, posted tweets were classified into topics related to trip purpose namely: visiting family and friends, vacation and leisure, going to church or religious activities, shopping, restaurant, and café, and going to work or school. The major origin and destination provinces of travelling X/Twitter users was determined and further exploratory analysis showed the influence of seasonal events on the X/Twitter intensity of travelling and their purpose of travel. The study demonstrates the potential use of social media data in studying travel behavior in the Philippines which can complement traditional national surveys on domestic travel.
Source or Periodical Title
UP Los Baños Journal
Volume
21
Issue
1
Page
102-118
Document Type
Article
College
College of Arts and Sciences (CAS)
Language
English
Subject
Geo-located tweets; Latent Dirichlet allocation model, Travel behavior analysis
Recommended Citation
Roldan, Ronald R. Jr. and Albacea, Zita VJ., "Classifying geo-located X/Twitter data using probabilistic topic modelling to analyze domestic travel patterns in the Philippines" (2023). Journal Article. 5896.
https://www.ukdr.uplb.edu.ph/journal-articles/5896