A simple Django website that allows people to annotate which crawled tweets that supposedly are about current COVID-19 symptoms are actually describing them. The whole annotation pipeline is supposed to work out in the open: Participants can
- Download all the annotations being done by the crowd
- Get live updates on how the annotations change the predictions and correlations to COVID-19 symptom reports
- Access the original notebooks for the analyses and run them directly in MyBinder.
Run the following (assuming you already have pipenv and heroku installed)
pipenv install
pipenv shell
heroku local:run python manage.py migrate
heroku local:run python manage.py import_tweets
heroku local
Your website should be up at 127.0.0.1:5000
.
The site is already setup to work multi-lingual. If you change any of the template texts you can get create updated translation files with
heroku local:run python manage.py makemessages
If you use Transifex for doing the translation do
tx push -s
to push the latest source language (english). Once the translations are done you can get them and compile for usage via:
tx pull
heroku local:run python manage.py compilemessages