Media

Media

Problem

Customer Sentiment Analysis

The client belongs to the media and entertainment industry, where a lot of web content is generated on a regular basis. In order to improve the content quality and make it more specific to the customer’s needs, it becomes essential to understand the customer sentiment with respect to any particular content being broadcasted. Hence, the client wants to perform customer sentiment analysis by interpreting reviews, social media tweets and posts which would help in identifying the percentage of positive and negative feedback.

Technology Used

Python, NLTK, Beautiful Soup (for web scraping posts, tweets and reviews)

Solution

Since the problem involves text from social media posts and tweets, natural language processing becomes an automatic choice of textual analysis. The NLP model is able to classify the posts, messages and conversation fragments by the sentiment they express defining the emotions hidden behind the context. This was achieved by using a specific library NLTK in Python with the help of an ensemble of three different modeling techniques, i.e., Logistic Regression, Stochastic Gradient Classifier and Support Vector Machines. The output of these models classified the emotion in the review posts as positive, negative or neutral, which helped the client to further improve their web content in accordance with customer needs.

Contact Us

Call: 020 3573 6917

Email: enquiries@bsbiconsulting

4 + 3 =