Understanding Weekly COVID-19 Concerns through Dynamic Content-Specific LDA Topic Modeling.

A set of 40 covid-related topics (per week) derived from Twitter starting on March 12, 2020.

Read the full publication here.

Topics

Each week's topics are available in csv format.

week_*_cp

• id: auto-incremented numeric row id
• term: unigram in topic
• category: Numeric topic id (from 0 to 39)
• weight: Conditional probability of the topic given the unigram, as derived through the LDA process.

week_*_loglik

• id: auto-incremented numeric row id
• term: unigram in topic
• category: Numeric topic id (from 0 to 39)
• weight: Posterior likelihood

Dates

• Week 1: begins March 12, 2020
• Week 2: begins March 19, 2020
• Week 3: begins March 26, 2020
• Week 4: begins April 02, 2020
• Week 5: begins April 09, 2020

Citation

Please cite the following paper if you use this data.

@inproceedings{zamani2020understanding,
    title = "Understanding Weekly COVID-19 Concerns through Dynamic Content-Specific LDA Topic Modeling",
    author = "Zamani, Mohammadzaman  and
      Schwartz, H. Andrew  and
      Eichstaedt, Johannes  and
      Guntuku, Sharath Chandra  and
      Ganesan, Adithya Virinchipuram  and
      Clouston, Sean  and
      Giorgi, Salvatore",
    booktitle = "Proceedings of the Fourth Workshop on Natural Language Processing and Computational Social Science",
    year = "2020",
    publisher = "Association for Computational Linguistics",
}

Contact

Please contact mzamani [at] cs [dot] stonybrook [dot] edu with any questions.

License

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