MULTILINGUAL DETECTION AND MAPPING OF EMERGENCY AND DISASTER-RELATED TWEETS

The Philippines is considered as a disaster-prone country in Southeast Asia. Today, social media such as twitter serves as a communication outlet and majority of the post are written in English. This is a problem or gap to those who are not well-versed in a foreign language or cannot even read or understand English. This study promotes the use of local language by translating the keyword using the specified language of the identified region. It will enhance and bridge the gap between the major speaking language from the local areas of the country specifically in the Ilocos Region. The tool will search disaster and emergency-related keywords in local language for extraction. Social network's API and tools will be used for community detection and extraction of data. This shall analyze the properties of the community structure detected from Filipino social media users who posted about the disaster in the local language. This study will determine the geolocation and community structure of the disaster and emergency-related post based on the tweet's coordinates, and analyze community structure formed and compare it to actual patterns of disaster-affected areas. Maps will be utilized as a crowdsource to identify the disaster and emergency-related tweets in Ilocos Region. This will also improve the development and use of the tool through the multilingual Twitter data and in real-time detection of disaster-related tweets so that appropriate action may be done promptly. This paper presents the possibility of the affected community which gives bigger changes of possible projecting the exact location for a reliable report to the government for a faster response.