WhatsApp Chat Sentiment Analysis Using BERT

Overview

This project aims to analyze sentiments in WhatsApp chat exports using BERT (Bidirectional Encoder Representations from Transformers). By uploading a chat file, the system preprocesses the messages, applies various visualization techniques, and performs sentiment analysis on each message, providing insightful results.

Features

• Chat File Upload: Users can upload their WhatsApp chat exports for analysis.
• Preprocessing: Messages are cleaned and normalized for analysis.
• Visualization: Various techniques to visualize text data, including word clouds and bar charts.
• Sentiment Analysis: Uses BERT to perform sentiment analysis on each message and present the results.

Modules

Preprocessor

The preprocessor.py module is responsible for preparing the chat data for analysis. It includes functions for:

• Data Cleaning: Removing unnecessary characters and noise from the text.
• Normalization: Standardizing the text data for consistent analysis.
• Message Segmentation: Splitting the chat data into individual messages with timestamps and user information.

Key Functions

• get_sentiment_score(message): Uses BERT to get the sentiment label and score for a given message.

Helper

The helper.py module contains various functions for data analysis and visualization. It includes:

• Statistical Analysis: Functions to calculate total messages, total words, and other statistics.
• Visualization: Functions to create word clouds, bar charts, and other visualizations of the chat data.

Key Functions

• fetch_stats(selected_user, df): Returns the total number of messages, words, media messages, and links.
• most_busy_users(df): Identifies the most active users in the chat.
• create_wordcloud(selected_user, df): Generates a word cloud for the most used words.
• most_common_words(df, selected_user): Identifies the most common words used in the chat.
• emoji_helper(selected_user, df): Analyzes the usage of emojis in the chat.
• monthly_timeline(selected_user, df): Generates a monthly timeline of messages.
• daily_timeline(selected_user, df): Generates a daily timeline of messages.
• week_activity_map(selected_user, df): Analyzes weekly activity patterns.
• month_activity_map(selected_user, df): Analyzes monthly activity patterns.
• activity_heatmap(selected_user, df): Creates a heatmap of activity by day and hour.
• sentiment_analysis(selected_user, df): Analyzes the sentiment distribution in the chat.
• sentiment_analysis_by_user(selected_user, df): Analyzes sentiment by user.

App

The app.py module is the main entry point for the Streamlit application. It provides the user interface for uploading chat files, displaying visualizations, and presenting sentiment analysis results.

Key Features

• File Upload: Allows users to upload WhatsApp chat exports.
• Interactive Visualizations: Displays various visualizations of the chat data.
• Sentiment Analysis Results: Presents sentiment analysis results using BERT.

Running the Application

To run the application, follow these steps:

1. Install Dependencies: Make sure you have Python and the required packages installed.

   pip install -r requirements.txt