README.md

Covid Mask Sentiment

Contributing to the project

• Make sure pre-commit hooks are installed by running pre-commit install in the root directory (one time set up).
• Obtain Twitter developer credentials here.
• Database URL = <dialect>+<driver>://<user>:<password>@host:port/database
• Create and fill secrets file using cp .envrc.example .envrc
• Give direnv permission to read .envrc by running direnv allow
• Install all dependencies using pip install -r requirements.txt

Start the Docker container by running sh ./start_app.sh.

Annotation

A custom annotation tool was used to label the 2,500 tweets as positive, negative or neutral.

• Positive: wearing masks during the pandemic is helpful in diminishing the spread of covid-19.

• Negative: wearing masks is ineffective or harmful for the spread of covid-19

• Neutral: tweet mentions key words but does not express any sentiment.
  e.g. poses a question, product marketing, quoting any text or reference without adding an opinion, sentence unintelligible, etc.

Note: A small number of tweets were dropped from the analysis if the sentence was unintelligible or contained languages other than English that made sentiment impossible to assign.