[4G1-GS-4] Web intelligence:

[4G1-GS-4-03] Japanese atmosphere index measured from social media data

Suzuka Matsumoto¹, 〇Akito Nakanishi¹, Naofumi Otomo¹, Koutarou Tamura², Junichi Shiozaki², Yukie Sano¹ (1. University of Tsukuba, 2. Nomura Research Institute, Ltd.)

Keywords:SNS, Sentiment analysis, Time series

We newly developed index, called the “Atmosphere Index,” to capture the collective emotions of Japanese people with seven emotional states, including fatigue. This was achieved by utilizing a POMS-based emotion dictionary applied to Twitter (now X) data based on the previous research. The index showed distinct fluctuation patterns in emotions, including long-term trends, annual and weekly periodicities, using the Prophet model. This study highlights a new approach in collective emotion analysis using Japanese data with diverse emotions. Our findings are significant for understanding societal emotional responses, with potential applications in research on events like the COVID-19 pandemic. Additionally, our results are invaluable for companies in crafting marketing and public relations strategies by gauging consumer emotional trends and needs.