このアイテムのアクセス数: 57
このアイテムのファイル:
ファイル | 記述 | サイズ | フォーマット | |
---|---|---|---|---|
s42421-023-00080-z.pdf | 878.8 kB | Adobe PDF | 見る/開く |
タイトル: | On the Relationship Between Crowdsourced Sentiments and Mobility Trends During COVID-19: A Case Study of Kyoto |
著者: | Sun, Wenzhe ![]() ![]() Kobayashi, Hironori Nakao, Satoshi ![]() ![]() ![]() Schmöcker, Jan-Dirk |
著者名の別形: | 孫, 文哲 小林, 弘典 中尾, 聡史 |
キーワード: | Tweet sentiment Emotion identification Mobility and activity COVID-₁₉ Crowdsourced data |
発行日: | Dec-2023 |
出版者: | Springer Nature |
誌名: | Data Science for Transportation |
巻: | 5 |
号: | 3 |
論文番号: | 17 |
抄録: | COVID-19 has significantly changed people’s attitudes and behavior in Japan. In this study, we firstly conducted a sentiment analysis on the contents of tweets posted on Twitter to understand the changes in people’s emotions due to the COVID-19 pandemic and related restriction policies. ML-ask, an open-source analysis system for textual input in Japanese, is used for this sentiment analysis. Its emotion identification is lexicon-based and can identify ten different emotions including joy, sadness, dislike, and anger. Secondly, we investigated the impacts of such “crowdsourced” sentiments on mobility and activity participation by regression models. Publicly available, also crowdsourced, Google’s COVID-19 Community Mobility Reports are used as dependent variables, and the identified tweet sentiments are used as independent variables together with other conventional variables such as dummies for policies, weekends, and holidays. As a result, it was confirmed that the changes in mobility and activity participation could be explained more accurately by the counts of tweet sentiments. The model fit was significantly improved if sentiments are added as variables, compared to the cases in which only the number of tweets or the number of daily new cases was used. In addition, we test the effect of time lag. More specifically, the model fit was higher when using the averaged emotion counts of the past 7 days than using the count of emotions on the same day. We conclude that using tweet sentiments offers the feasibility to improve the prediction levels of urban mobility. |
著作権等: | This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1007/s42421-023-00080-z The full-text file will be made open to the public on 12 August 2024 in accordance with publisher's 'Terms and Conditions for Self-Archiving'. This is not the published version. Please cite only the published version. この論文は出版社版でありません。引用の際には出版社版をご確認ご利用ください。 |
URI: | http://hdl.handle.net/2433/285764 |
DOI(出版社版): | 10.1007/s42421-023-00080-z |
出現コレクション: | 学術雑誌掲載論文等 |

このリポジトリに保管されているアイテムはすべて著作権により保護されています。