Analyzing Geo-tagged Tweets about COVID-19 in Japan using MACD-Histogram-based Burst Detection

The analysis of real-world information with big data is an important task. However, it is difficult to determine the attention received by particular topics in certain regions. This study attempts to analyze COVID-19-related geo-tagged tweets in Japan. A method is proposed based on the moving average convergence/divergence (MACD)-histogram-based burst detection technique. Moreover, the proposed method utilizes heatmaps to visualize the burstiness. The proposed method can detect the amount of attention received by a topic along with its regional differences. We performed experiments using geo-tagged tweets related to COVID-19 in Japan and analyzed the obtained results.