Topic Chosen: Email sentiment analysis. We will be creating a process for enterprises to quickly analyze customers’ sentiment about their product/ service.
Problem Statement: Every customer is important to an enterprise for the business growth. To keep the customers loyal to your product/ service, we need to address their issues/ concerns on priority. By building a system which can analyze thousands of emails received every day and categorize emails based on their sentiments (e.g., customers unhappy about a certain product/ service).
Relevance to the Course Subject: To achieve this goal, we will have to perform Text Mining and Analysis. Strategy:
Dataset Selection: We have chosen a dataset available in Kaggle which is a collection of customer complaints from various industries.
Dataset Source: Evaluating Customer Service https://www.kaggle.com/janiobachmann/evaluating-customer-service-sentiment-analysis/data
Algorithms/ Techniques: We are considering using a Convolutional Neural Network algorithm.
Source Code: email_data_cleaning.ipynb and Sentiment_CNN.ipynb
Documentation: Documentation