Extraction of Urban Keywords through Twitter Data Analysis

Abstract

We analyzed the keywords of the city using SNS, and examined whether it is possible to extract the requests from the citizens who have not been able to grasp them before. We collected and analyzed Tweets from five cities, and extracted characteristic words. We collected text data of about 3 million Tweets in a year. On the other hand, since there are many words that are unrelated to the target area, we formulated rules and removed unnecessary Tweets. Furthermore, a hierarchical cluster analysis identified irrelevant clusters and removed them. The extracted words were subjected to correspondence analysis to identify the time of submission and the characteristics of the region to obtain the broad keywords of city planning.