A set of 50 pregnancy-related topics derived from a 10% Twitter sample from 2009-2015.

Read the full publication here.

Available in both csv format and as a MySQL dump.

• id: auto-incremented numeric row id
• term: unigram in topic
• category: Numeric topic id (from 0 to 49)
• weight: Conditional probability of the topic given the unigram, as derived through the LDA process.

• id: auto-incremented numeric row id
• term: unigram in topic
• category: Numeric topic id (from 0 to 49)
• weight: Posterior likelihood

Please cite the following paper if you use this data.

@inproceedings{abebegiorgi2020quantifying,
    title={Quantifying Community Characteristics of Maternal Mortality Using Social Media}, 
    author={Abebe, Rediet and Giorgi, Salvatore and Tedijanto, Anna and Buffone, Anneke and Schwartz, H. Andrew}, 
    booktitle={The World Wide Web Conference},
    year={2020}
}

Licensed under a GNU General Public License v3 (GPLv3).