Corpus Analysis of Earthquake Related Tweets through Topic Modelling

Abstract

The Philippine archipelago is one of the most disaster-prone area due to its location. As Twitter grows everyday, it has now become a valuable source of people's opinion. The main purpose of the study is to use Twitter, as text corpora in the attainment of disaster risk reduction. Earthquake related tweets posted from July 1, 2017 to August 31, 2017 were programmatically collected. Data cleaning was made by removing noisy words, hence, from 90,692 collected tweets this resulted to 41,500 cleaned tweets. Topic modeling identifies patterns in a corpus. Findings revealed that there is a need to strengthen the conduct of earthquake drill and early warning notices to lessen the vulnerability of people and property, and reduce the causal damages of earthquake, thus improve its community resiliency. Topic results shall be discussed with concerned agencies for its possible consideration and inclusion to enhance the earthquake-related disaster management plan.

Information Preference and Information Supply Efficiency Evaluation before, during, and after an Earthquake: Evidence from Songyuan, China

“…Field investigation revealed disaster information needs in detail, but the timeliness and regional range of acquiring disaster information were limited [ 82 ]. Future methods of acquiring data may consider combining field investigation with big data mining, such as using data extracted from social media [ 24 , 83 , 84 , 85 , 86 , 87 , 88 , 89 , 90 , 91 ], Google Trends [ 92 , 93 ], and Baidu Index [ 94 , 95 ]. In particular, these several data sources can make it possible to quite easily track the behavior of Internet users in real time.…”

Tweet Trends and Emotional Reactions: A Comprehensive Sentiment Analysis of Twitter Responses to the Earthquake in Turkey and Syria

“…In [19], the natural language processing (NLP) analysis method was applied to analyze the temporal and spatial sequence characteristics of the text and comments contents on microblogs posted by people as well as official media outlets within 6 days after the 2022 Lushan M6.1, Maerkang M5.8 and Luding M6.8 earthquakes. Earthquake-related tweets from one of the most disaster-prone areas, the Philippine archipelago, posted from July-August of 2017 were analyzed in the article [20] using topic modeling to identify patterns in a corpus. While using text data provided by the Euro-Mediterranean Seismological Centre (EMSC) for the Aegean Earthquake, the authors in [21] presented a sentiment and topic analysis according to the intensities reported by their users in the Modified Mercalli Intensity (MMI) scale using supervised classification and calculated statistical relationships between intensities reported in the MMI by the app users, polarities, and topics addressed in their comments.…”

Discovering Insights via Hybrid Thematic Analysis: A Case Study on Disaster Risk Reduction and Management for Legazpi City, Albay

“…The study highlights that Filipino Twitter users shared various information related to the disaster, including updates on the disaster's progression, personal experiences, and calls for help. Similarly, in the works of [12] and [13], the authors performed NLP-based techniques like topic modeling and sentiment analysis on typhoons and earthquake-related tweets in the Philippines. Another NLP-based paper [14], utilizes 976 suggestions on how their village can help them better prepare for a disaster implemented computational methods, specifically topic modeling and word2vec, to assist in the analysis of qualitative data on disaster risk reduction suggestions.…”