Data analystics for earthquake social tweets using clustering approach

• Main Author: Wang, Lanlan
• Other Authors: Xiao Gaoxi
• Format: Theses and Dissertations
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/68675
• Institution: Nanyang Technological University

Due to natural disasters cause great losses to human beings, giving assistances to earthquake-stricken areas blindly is a waste of resources. This thesis aims to study earthquake social tweets data for understanding the patterns of people's concerns on the ground through clustering approaches. We clustered earthquake tweets into three clusters according to their metrics. Government could take the appropriate measures according to different kind of earthquakes. This thesis is based on FCM clustering algorithm, but compare with K-means clustering algorithm. Experimental results show that the samples in clusters obtained from FCM clustering algorithm is more consistent when there are more than two or three variables features. FCM clustering algorithm is more flexible when dealing with practical problems. These earthquakes are divided into 3 clusters. Earthquakes in cluster 1 located in areas where almost no people live or the earthquake magnitudes are small. People focus little on these earthquakes. Earthquakes in cluster 2 mainly located in high-density population areas, and earthquake magnitudes are generally ML5 to ML6. People attach close attention to an earthquake when it just occured, but this attention fades with time and gets stronger again due to certain reasons. Governments should preach basic earthquake knowledge to public and do some earthquake preventive measure education. Earthquakes in cluster 3, with earthquake magnitudes mainly ML6 to ML7, are mainly located in higher-density population areas. People living in these areas pay a lot of attention on earthquakes. Although tweets number also has fluctuations, the number of tweets keeps an increasing trend which shows that people keep concerning them. Governments need to pay their most attention to these places. In terms of areas having earthquakes in cluster 3, in addition to preaching seismic knowledge and regular rehearsing, governments also need to do assistance measures to reassure and pacify the public, like psychological appease and economic assistance.