Does what it says. Monitor specifics hashtags or brand names. Just an exercise atm.
It will heavily rely on docker images for most of the work cause I'm a lazy ass.
- Define Brand (). Store hashtags for a brand (or any other denominator)
- Collect Response for the hashtag on a serie of different platforms:
- Automatise it (airflow / papermill)
- store it in an SQL database. ( Maybe elastic could prove efficient. Dunno, depending on volume.)
- Transform
- langage detection
- sentiment analysis
- normalize interactions
- anlyze Images ? (later)
- Aggregate and score the whole thing
- Display it (jupyter then flask)
- Add Flask Cache for rquests.
#LATER
- Collect Response for the hashtag on a serie of different platforms:
- Google Alerts
- Add Roles to the app
- Postgresql
- airflow
- celery
- redis
- flask
- jupyter