The twitter sentiment platform has three main components, the collector, the processor, and the API.
The collector utilizes Twitter's stream to collect firehose tweets filtered by keyword.
We use the great Tweepy library as the platform's foundation.
The processor conducts sentiment analysis on a batch of tweets and writes them to the database.
Sentiment is scored using Vader Sentiment Analysis implemented in NLTK.
The sentiment API serves tweet polarity to the frontend.
The API is built on Flask and connects to a SQLite3 database.
The sentiment Bot publishes stats to Twitter on a cron.
Tweepy is used to handle publishing.
This platform is built on Python3. This assumes you have a Python3 environment enabled.
First, run pip install -r requirements.txt to install the dependencies. You will also
need some corpora from NLTK, so run nlkt.download().
Once you have the dependencies, copy config.ini.example to config.ini and update as
appropriate.
Next, create a SQLite3 database using the schema provided in databases/.
After that, follow the READMEs in collector/, processor/, and api/ to run
the components.