Tweets containing "climate change" with topic annotations

This dataset contains the Twitter IDs of all ~20M tweets containing the phrase "climate change" 2018-2021. Additionally, it contains the topical annotations and 2D semantic representation of our thematic analysis based on ~980 topic clusters that are grouped by hand into seven themes (COVID-19, Politics, Contrarian, Movements, Solutions, Impacts, Causes) as well as "non-relevant/spam", "others", and highlighting of potentially interesting topics. Code and additional notes are available on GitHub: https://github.com/TimRepke/twitter-climate The topics, including statistics and the annotator labels for broader themes (aka "super topics") are contained in the spreadsheet. This data is extrapolated to the tweets contained in the share.jsonl file containing one json object per line with the following fields: 'rel': true iff Tweet is contained in analysis 'filters': null if Tweet is not included, otherwise contains an object with "reasons" why this tweet was excluded 'dup': 1 iff this is a duplicate (excl first) 'lan':  1 iff language is English (and not None) 'txt': 1 iff status text is not None 'mit': 1 iff text has minimum number of tokens (>=4) 'mah': 1 iff text has less than maximum number of hashtags (<=5), 'pfd': 1 iff tweet was posted after 01.01.2018 'ptd': 1 iff tweet was posted before 31.12.2021 'cli': 1 iff tweet actually contains "climate change" (API matches some false positives) 'ann': null if Tweet is not included, otherwise contains an object with topic annotations 't_km':  topic (based on "keep & majority vote" strategy) 't_kp':  topic (based on "keep & closest topic centroid [proximity]" strategy) 't_fm':  topic (based on "drop sample topic [fresh] & majority vote" strategy) 't_fp':  topic (based on "drop sample topic [fresh] & closest topic centroid [proximity]") 'st_int':  theme annotation "Interesting" 'st_nr':  theme annotation "Non-relevant / spam" 'st_cov':  theme annotation "COVID" 'st_pol':  theme annotation "Politics" 'st_mov':  theme annotation "Movements" 'st_imp':  theme annotation "Impacts" 'st_cau':  theme annotation "Causes" 'st_sol':  theme annotation "Solutions" 'st_con':  theme annotation "Contrarian" 'st_oth':  theme annotation "Other" 'x':  x position in 2D representation 'y':  x position in 2D representation 'sample':  true iff this tweet was in the original topic model sample