v1.0 data

See the code used that analyzes the data in this repository at wwbp/robust_spatiotemp

Contains the generated 2020 mental health scores generated in Robust language-based mental health assessments in time and space through social media. These scores control for 2019 findings.

• lbmha_yw_cnty.csv: depression and anxiety scores for 2020 measured from Twitter data. This data controls for the 2019 data, so 2019 is not included in this file.
  • id unique row identifier [0, 1, 2, ...]
  • yearweek_cnty concatenation of the year, week number (ISO format), and county (fips code) feat: either DEP_SCORE (depression) or ANX_SCORE (anxiety), derived from an adapted lexicon [2019_01:36091, 2020_52:22069]
  • score_category Depression (DEP_SCORE) or Anxiety (ANX_SCORE)
  • n_lexwords total count of words signalling the feat from all users in that yearweek_cnty
  • score weighted average of scores signalling the feat from all users in that yearweek_cnty, weights are meant to better align with actual demographics of population and only occurs when sufficient users are present
  • score_unweighted average of scores signalling the feat from all users in that yearweek_cnty are present
  • score_std standard deviation of scores signalling the feat from all users in that yearweek_cnty are present
  • moderate_reliability_criteria whether the score is considered reliable based on the criteria of UT 50 [1 if true; 0 if false]
  • high_reliability_criteria whether the score is considered reliable based on the criteria of UT 200 [1 if true; 0 if false]
  • n_users number of unique users measured in that yearweek_cnty
  • cnty redundant column for easy sorting
  • county_name expanded name of the county
  • state_abbr standard state 2-letter abbreviation of state
  • date The Wednesday of the ISO week
  • date_range Monday-Sunday of the ISO weeks
  • yearweek redundant column for easy sorting
• lbmha_yw_cnty_undifferenced.csv depression and anxiety scores for 2019 and 2020 measured from Twitter data. This data does not include any correction for other years. The data format is shared from lbmha_yw_cnty.csv.