PC Activity Analyzer

📌 Overview

PC Activity Analyzer is a privacy-first digital habit tracking tool that monitors your PC usage and provides AI-powered insights. All data processing and AI analysis happens locally on your device - no cloud uploads, no external API calls.

---

✨ Key Features

• Real-time Activity Tracking - Monitors active window usage on your PC
• Interactive Dashboards - Visualize your digital habits with charts and analytics
• AI-Powered Insights - Natural language chatbot answers questions about your usage patterns
• Smart Filtering - Filter data by app, date, time, and custom criteria
• Privacy-First - All data and AI processing stays on your device
• Data Export - Export your usage data as CSV for external analysis

Dashboard & Analytics

Detailed analytics with hourly, daily, and weekly breakdowns

Selected App – Avg. Session Duration & Analytics

AI-Powered Chat

Natural language chatbot for querying your usage patterns

Settings

---

🏗️ Architecture

Core Components

• Active Window Tracker: Background service that logs app usage to CSV
• Dashboard & Analytics: Streamlit-based interactive UI for data visualization
• RAG Chatbot: AI assistant that answers questions using your own usage data
• ChromaDB: Local vector database for fast semantic search
• Ollama + LLM: Local large language model for generating AI responses

Data Pipeline

graph TD
    A[Active Window Tracker] --> B[usage_log.csv]
    B --> C[Streamlit App]
    C --> D[Dashboard/Analytics]
    C --> E[RAG Chatbot]
    E --> F[ChromaDB Vector DB]
    E --> G[Ollama + LLM]

Loading

How It Works

1. Data Collection - Background tracker monitors active windows and logs usage to CSV
2. Data Processing - Raw data is cleaned and enriched with time-based features
3. Embedding & Vectorization - Usage summaries are converted to vector embeddings
4. AI-Powered Insights - RAG pipeline retrieves relevant data and generates natural language responses
5. Interactive Visualization - Streamlit UI presents dashboards and chatbot interface

---

🔧 Technical Stack

• Python 3.11+ - Core programming language
• Streamlit - Web UI framework for dashboards
• LangChain - RAG pipeline orchestration
• Ollama + LLaMA3 - Local large language model
• ChromaDB - Vector database for embeddings
• Sentence Transformers - Text embedding generation
• pywin32 / win32gui - Windows API for activity tracking
• Plotly - Interactive data visualizations
• Pandas - Data processing and analysis

---

📁 Project Structure

pc_activity_analyzer/
├── app.py                      # Main Streamlit application
├── active_windows_tracker.py   # Windows activity logging service
├── analyze_usage.py            # Standalone data analysis script
├── requirements.txt            # Python dependencies
├── README.md                   # Project documentation
├── styles.css                  # Custom Streamlit styles
├── data/
│   └── usage_log.csv           # Generated usage data
├── chroma_db/
│   └── chroma.sqlite3          # ChromaDB vector database
├── components/
│   ├── ai_chat.py              # AI chatbot UI component
│   ├── analytics.py            # Analytics dashboard component
│   ├── dashboard.py            # Main dashboard component
│   └── settings.py             # Settings and configuration UI
├── chatbot/
│   ├── data_loader.py          # Data loading utilities
│   ├── data_processing.py      # Data processing helpers
│   ├── document_processor.py   # Document creation for RAG
│   ├── embedding_store.py      # Embedding storage management
│   ├── llm_handler.py          # Ollama LLM integration
│   ├── quick_analysis.py       # Quick analytics processing
│   ├── quick_questions.py      # Predefined question patterns
│   ├── usage_data_rag.py       # RAG pipeline implementation
│   └── vector_store_manager.py # ChromaDB management
├── utils/
│   ├── data_loader.py          # General data loading utilities
│   ├── helpers.py              # Common helper functions
│   └── instance.py             # Singleton pattern helpers
└── images/                     # Screenshots and documentation assets

---

🛠️ Installation

1. Clone the Repository

git clone <repo-url>
cd pc-activity-analyzer

2. Install Python Dependencies

pip install -r requirements.txt

3. Install and Setup Ollama

Download and install Ollama from ollama.com, then:

ollama serve
ollama pull llama3

4. Generate Initial Data

Run the activity tracker to start collecting usage data:

python active_windows_tracker.py

Let it run for a while to collect some initial data, then stop it with Ctrl+C.

---

⚡ Quick Start

1. Start Ollama Service

ollama serve

2. Launch the Application

streamlit run app.py

3. Access the Dashboard

Open your browser and navigate to http://localhost:8501

4. Explore Features

• Dashboard - View your usage statistics and trends
• Analytics - Dive deeper into your digital habits
• AI Chat - Ask questions about your usage patterns
• Settings - Configure filters and export options

---

⚙️ Configuration

Data Sources

• Usage Log: data/usage_log.csv (auto-generated by tracker)
• Vector Database: chroma_db/ (created automatically)

Customization Options

• Ollama Model: Default is llama3 (configurable in code)
• Tracking Interval: Adjustable in active_windows_tracker.py
• Data Retention: Configure how long to keep historical data
• UI Themes: Custom styles in styles.css

---

🚀 Performance Features

• Efficient Caching - Embeddings and processed data are cached for fast access
• Vector Search - ChromaDB provides millisecond semantic search
• Streamlit Optimization - Smart caching reduces data processing overhead
• Local Processing - No network latency from cloud API calls

---

🔒 Privacy & Security

Complete Data Privacy

• Local Storage Only - All data remains on your device
• No Cloud Uploads - Zero external data transmission
• Local AI Processing - LLM runs entirely on your hardware
• Open Source - Full transparency of data handling

Data Control

• Export Options - Full control over your data export
• Deletion Controls - Easy data cleanup and removal
• Access Logs - Track what data is being accessed

---

📄 License

This project is licensed under the MIT License – see the LICENSE file for details.

---

🙏 Acknowledgments

Inspired by the need for local, privacy-first digital habit analytics and AI-powered insights.