L0veMathur · Nikkyzzzz · Oct 25, 2025
diff --git a/README.md b/README.md
@@ -1,3 +1,195 @@
-Input your exported WhatsApp chat in the "WhatsappChat to csv.py" to make a cleaned .csv file from raw .txt file. Use the .csv file as input for "ChatAnalyzer.py" to get your chat analyzed on various metrics. It also has integrated Cohere chat model to ask questions about the chat data and to get certain insights
 
-Remember to replace your trial cohere api keys with "your-api-keys-here". The keys are free and can be generated on cohere website. Sign-up for Cohere and get your free keys generated. After signing up for Cohere go to https://dashboard.cohere.com/api-keys adn generate a trial key.
+# 💬 WhatsApp Chat Analyzer
+
+A **Streamlit-based Machine Learning Dashboard** that analyzes WhatsApp chats to extract deep conversational insights — including message statistics, user behavior, sentiment, and predictive analytics.  
+The app integrates **Cohere’s NLP API** to provide an **AI-powered chatbot** for dynamic insights and contextual question answering.
+
+---
+
+## 🚀 Key Features
+
+### 📊 Chat Data Analysis
+- Parses and cleans WhatsApp exported chat text into a structured CSV/Excel file.  
+- Displays total messages, active days, average messages/day, and sender activity.
+
+### 🧠 Machine Learning Insights
+- Trains ML models (using **Scikit-learn**) to:
+  - Predict message reply times.
+  - Identify probable “love” or “regular” messages.
+  - Analyze message sentiment using **TextBlob**.
+
+### 🤖 Cohere AI Chatbot
+- Integrates **Cohere’s “command-light”** LLM model to interpret chat data.
+- Users can ask natural language questions about their chats, e.g.:
+  - “Who talks more during weekends?”
+  - “When are we most active?”
+  - “What’s the tone of our conversations?”
+
+### 📈 Visualization & Reports
+- Interactive **Streamlit dashboard** with:
+  - Word clouds
+  - Time-based message frequency plots
+  - Message length distributions
+  - Keyword trends
+- Generates a full **analytical report** summarizing behavioral insights.
+
+---
+
+## 🧩 Tech Stack
+
+| Component | Technology |
+|------------|-------------|
+| Frontend | Streamlit |
+| Backend | Python |
+| Data Processing | Pandas, NumPy, re, openpyxl, BytesIO |
+| Machine Learning | scikit-learn, TextBlob |
+| Visualization | Matplotlib, Seaborn, WordCloud |
+| NLP & LLM Integration | Cohere API (Command-Light model) |
+
+---
+
+## 🧠 Project Architecture
+
+```
+
+WhatsApp Chat Analyzer
+│
+├── 📄 WhatsAppChat to csv.py      # Converts raw WhatsApp .txt to structured CSV/Excel
+├── 📄 ChatAnalyzer.py             # Streamlit dashboard with ML and Cohere integration
+│
+├── 📊 cleaned_messages.csv        # Output data after preprocessing
+│
+├── 📁 assets/                     # (Optional) images, wordclouds, icons
+│
+├── requirements.txt               # Python dependencies
+└── README.md
+```
+
+
+
+
+## ⚙️ Installation & Setup
+
+1. **Clone the repository**
+   ```bash
+   git clone https://github.com/<your-username>/whatsapp-chat-analyzer.git
+   cd whatsapp-chat-analyzer
+
+
+2. **Create a virtual environment**
+
+   ```bash
+   python -m venv venv
+   source venv/bin/activate       # For Linux/Mac
+   venv\Scripts\activate          # For Windows
+   ```
+
+3. **Install dependencies**
+
+   ```bash
+   pip install -r requirements.txt
+   ```
+
+4. **Add your Cohere API key**
+
+   * Open `ChatAnalyzer.py`
+   * Replace `"your-api-key-here"` with your **Cohere API key**
+     *(get it from [https://dashboard.cohere.com](https://dashboard.cohere.com))*
+
+5. **Run the Streamlit App**
+
+   ```bash
+   streamlit run ChatAnalyzer.py
+   ```
+
+---
+
+## 🧮 How It Works
+
+1. **Convert Chat → CSV**
+
+   * Use `WhatsAppChat to csv.py` to convert your exported WhatsApp `.txt` file.
+   * The script automatically extracts `Date`, `Time`, `Name`, and `Message` fields and saves:
+
+     * `cleaned_messages.csv`
+     * `cleaned_messages.xlsx`
+
+2. **Analyze in Dashboard**
+
+   * Upload or place the CSV in the same directory as `ChatAnalyzer.py`
+   * Run the Streamlit app to view dashboards such as:
+
+     * Message distribution
+     * Keyword frequency
+     * Sentiment trends
+     * Temporal activity (hourly, weekly)
+     * Word clouds and ML predictions
+
+3. **Chat with AI**
+
+   * Navigate to **AI Chat** section in sidebar.
+   * Ask natural questions about communication patterns — Cohere’s LLM generates insights in real time.
+
+---
+
+## 📊 Example Insights
+
+| Metric             | Example Output          |
+| ------------------ | ----------------------- |
+| Most Active User   | Alice                   |
+| Average Reply Time | 12.6 minutes            |
+| Common Words       | love, night, miss, good |
+| Peak Chat Hour     | 22:00 hrs               |
+| Sentiment Score    | +0.43 (Positive)        |
+
+---
+
+## 🧾 Sample Commands
+
+```bash
+# Convert WhatsApp chat to CSV
+python "WhatsAppChat to csv.py"
+
+# Launch the analyzer dashboard
+streamlit run ChatAnalyzer.py
+```
+
+---
+
+## 🧰 Dependencies
+
+```
+streamlit==1.39.0
+pandas==2.2.2
+numpy==1.26.4
+scikit-learn==1.5.2
+textblob==0.18.0.post0
+matplotlib==3.9.2
+seaborn==0.13.2
+wordcloud==1.9.3
+openpyxl==3.1.5
+cohere==5.5.8
+```
+
+---
+
+## 🧑‍💻 Author
+
+**Developed by:** *Love Mathur*
+**Email:** [your.email@example.com](mailto:your.email@example.com)
+**Cohere Model Used:** `command-light`
+**License:** MIT
+
+---
+
+## 🧾 Future Enhancements
+
+* Integration with OpenAI GPT models.
+* Advanced clustering of chat topics using sentence embeddings.
+* Sentiment-over-time trend visualization.
+* Real-time chat ingestion and dashboard auto-refresh.
+
+
+
+
+