Skillbot: A Conversational Chatbot based Data Mining and Sentiment Analysis

With the advent of technology, Artificial Intelligence is emerging exponentially. Using this advancement, chatbots are widely used in various sectors to accommodate users with their queries without waiting. In this study, work in the development, training, improvement, and chat sentiment analysis of Skillbot Chatbot is performed. First, the data was scrapped using tools like the GPT2 model from the Gov. UK website, and that data was used to train intents for the Skill bot model. After successful training, testing, and evaluation of Skillbot for better performance, conversations of users were analyzed deeply. Sentiment analysis was also performed as it is important to train the Skillbot to efficiently respond to users. Then, this project was deployed on Streamlit named Conversation Analyzer. Analysis was performed using different technologies like Natural Languages processing, Vader model for sentiment analysis, TextBlob for topic modeling of conversations, Streamlit for visualization, Rasa, Artificial Intelligence, and machine learning. Chatbot training with cleaned data and conversation analysis would be very beneficial for Skillbot to give users better services. The findings with massive data wrangling, model training for Skillbot, and chat analysis would provide results' evaluations with successful and unsuccessful dialogues with insights to help warrant future research and Skillbot improvement.