Harvard CGA Geotweet Sentiment Archive

Harvard CGA Geotweet Sentiment Archive is a subset of Harvard CGA Geotweet Archive v2.0 enriched with a sentiment score. It contains the tweet identification records along with a sentiment score based on tweet text for about 4.3 billion geo-tagged tweets since 2019. This sentiment score was calculated using Bidirectional Encoder Representations from Transformers. More information about this methodology can be found in our Nature Paper on Twitter Sentiment Geographical Index. This dataset is available to the academic community at large, unlike the Harvard CGA Geotweet Archive v2.0 which is under Twitter's redistribution policy restriction for public sharing. It could serve as cross-validation data for publications that used data from Harvard CGA Geotweet Archive v2.0 . If you are interested in accessing this archive, please fill out our Geotweet Request Form. Before requesting or receiving Tweet IDs, requestors must agree to Twitter's Terms of Service, Twitter's Privacy Policy, and Twitter's Developer Policy . Geotweets IDs data provided by CGA can only be used for not-for-profit research and academic purposes. Recipients may not share CGA provided Tweet IDs or content derived from them without written permission from the CGA. Citations: If you use the Geotweet Archive in your research please reference it: "Harvard CGA Geotweet IDs Archive". ======================================================== Schema of Geotweet Census Archive Field name____TYPE____Description message_id----TEXT----Tweet ID score ----FLOAT----BERT sentiment score