We can use our model in order to classify tweets in real time using Twitter’s API. In order to simplify things, we will make use of a very popular wrapper library for the API, tweepy (https://github.com/tweepy/tweepy). Installation is easily achieved with pip install tweepy. The first step to accessing Twitter programmatically is to generate relevant credentials. This is achieved by navigating to https://apps.twitter.com/ and selecting Create an app. The application process is straightforward and should be accepted quickly.
Using tweepy's StreamListener, we will define a class that listens for incoming tweets, and as soon as they arrive, it classifies them and prints the original text and predicted polarity. First, we will load the required libraries. As a classifier, we will utilize the voting ensemble we trained earlier. First, we...
