Skip to content

Citation Compass

Explore citation networks, predict research connections, and analyze scholarly impact—powered by machine learning and graph analytics.

Get Started Live Demo GitHub

📚 Documentation Sections

  • Home

    Overview, quick start, and navigation to all documentation

  • Getting Started

    Installation, configuration, and your first analysis in under 10 minutes

  • User Guide

    Complete walkthrough of features, workflows, and best practices

  • Research Notebooks

    4-notebook pipeline for model training and comprehensive analysis

  • Developer Guide

    Architecture, API reference, and technical design decisions

  • Resources

    Practical guides for Neo4j monitoring and database maintenance

⚡ Quick Start

Get running in 3 steps:

git clone https://github.com/dagny099/citation-compass.git
cd citation-compass
pip install -e ".[all]"

streamlit run app.py
# No database needed! Load demo datasets from sidebar

cp .env.example .env
# Add Neo4j credentials for your own data
# NEO4J_URI=neo4j+s://your-database

Next: Full Getting Started Guide →

📊 Sample Workflows

Common Research Workflows

flowchart LR
    A["📄 Input Paper"] --> B["🧠 ML Analysis"]
    B --> C["🔮 Predictions"]
    C --> D["📊 Confidence Scores"]
    D --> E["📋 Reading List"]

    style A fill:#e3f2fd
    style B fill:#fff3e0
    style C fill:#e8f5e8
    style D fill:#fce4ec
    style E fill:#f1f8e9
1. Input a paper → Generate predictions → Validate with embeddings 2. Explore similar work → Build reading lists → Discover connections

flowchart LR
    A["👤 Select Author/Field"] --> B["🕸️ Build Network"]
    B --> C["🎯 Detect Communities"]
    C --> D["📈 Calculate Metrics"]
    D --> E["📄 Export Report"]

    style A fill:#ffebee
    style B fill:#e0f2f1
    style C fill:#f3e5f5
    style D fill:#e8f5e8
    style E fill:#fff3e0
1. Select author/field → Detect communities → Export LaTeX 2. Analyze collaborations → Identify key researchers → Track influence

flowchart LR
    A["📅 Date Range"] --> B["📊 Trend Analysis"]
    B --> C["📈 Growth Patterns"]
    C --> D["🔮 Predictions"]
    D --> E["📋 Insights"]

    style A fill:#f1f8e9
    style B fill:#e3f2fd
    style C fill:#fce4ec
    style D fill:#fff3e0
    style E fill:#e8f5e8
1. Choose date range → Analyze trends → Generate insights 2. Track paper impact → Monitor growth → Predict future citations

🎯 Platform Features

Citation Compass provides a comprehensive toolkit for citation network analysis, from data ingestion through final publication.

Core Capabilities

  • ML-Powered Predictions

    TransE embeddings learn paper relationships in vector space, enabling citation prediction with confidence scores and similarity rankings.

  • Network Analysis

    Advanced graph algorithms detect research communities, calculate centrality measures, and analyze temporal citation trends.

  • Interactive Visualization

    Clickable network graphs with real-time progress tracking make exploration intuitive—from initial data import to final insights.

  • Research Export

    Generate LaTeX tables, academic reports, and publication-ready visualizations in multiple formats (PDF, CSV, JSON).

Data Flow Architecture

The platform orchestrates four key pipelines: data ingestion from external APIs, ML training and prediction, network analysis, and interactive visualization. Each pipeline is optimized for its specific workload with caching and validation at every step.

Data Flow

🏗️ Architecture Overview

Citation Compass combines machine learning, graph analysis, and interactive visualization for academic citation networks.

System Architecture

Core Components:

  • Streamlit Dashboard - Interactive web interface with real-time visualizations

  • TransE ML Model - Citation prediction using knowledge graph embeddings

  • Neo4j Database - Graph storage optimized for citation network queries

  • Analytics Engine - Community detection, centrality measures, temporal analysis

🤝 Community & Support

Python Neo4j Streamlit License: MIT

Built for the research community 🔬