Authors: Associate Professor Ch. Neveen, Professor S.Rama Rao, Associate Professor M. Radhika
Abstract: It is difficult to organise massive volumes of data, assess sentiment accurately regardless of context, irony, or sarcasm, and create decent visualisations from unstructured chat data. This suggested research aims to develop a sophisticated conversation analyser specifically designed for WhatsApp discussions. It will make use of a variety of tools, including Python packages like pandas and matplotlib, as well as natural language processing algorithms. To better understand user dynamics, the computer use sentiment analysis to weed out unnecessary messages and aggregate warnings, allowing it to zero in on meaningful interactions. Developed on the Heroku cloud platform, this comprehensive analyser offers an intuitive application. It enables the display and statistical analysis of conversation data. By combining machine learning with natural language processing, the initiative improves users' comprehension of WhatsApp discussions, allowing them to easily extract crucial data and handle unwanted material. One method involves training a machine learning system to recognise patterns and emotions using VADER sentiment analysis and term frequency analysis. In terms of processing efficiency for user activity detection and sentiment analysis, the new technique outperforms previous models by 20%, while also increasing accuracy by 15%. Natural language processing (NLP), sentiment analysis, data visualisation, and insights into user behaviour are some of the topics that are often linked with Vader, Pandas, Matplotlib, and WhatsApp chat analyser.
