In this section, we will build our own model. There are existing datasets available that are related to Twitter feed data on the topic of product and movie reviews. You can pick a dataset that suits you; in this chapter, we will pick a dataset that has customer reviews.
A dataset that contains both positive and negative reviews of customers can be found at http://boston.lti.cs.cmu.edu/classes/95-865-K/HW/HW3/. You can download the dataset from the following link: http://boston.lti.cs.cmu.edu/classes/95-865-K/HW/HW3/epinions3.zip.
The aforementioned dataset has both positive and negative feedback about a product, as shown in the following screenshot:
We will train the dataset using the scikit-learn pipeline and LinearSVC. Let's take a closer look at both of these.