Afet sonrası sosyal medya verisini işleme

Social media has become a compelling concept in recent years. Contrary to old media, social media ensures bidirectional communication and sharing between sources and consumers. The most instant information shared by all users can be retrieved with a simple search. Although these platforms, where each person has a microphone, seem to be the right place to get instant information when important events occur, it creates chaos when everyone speaks simultaneously. Thus, it is essential to distinguish the desired information from this wild environment, especially in case of any disaster. Within the scope of this research, studies have been carried out for a smart disaster management system with the help of texts and locations shared on microblogging platforms after a disaster. For an application to conduct experimental research, tweets shared on Twitter during and after the Izmir earthquake happened on 30 October, 2020 were collected. The tweets were annotated as fully- and semi-supervised ways to create big and labeled datasets. Datasets are used for both content-based analyzes and classifications. Since the analysis and classification methods results are promising, it has been proven that the study can be useful in post-disaster scenarios.