Skip to content

README.md

Latest commit

 

History

History
228 lines (168 loc) · 7.63 KB

README.md

File metadata and controls

228 lines (168 loc) · 7.63 KB

SocialCollateral AI - Frontend Dashboard

Frontend interface for Jaringan Amanah (The Amanah Network), a Social Graph Engine built for the Amartha ecosystem. Provides interactive visualization and analysis tools for network-based risk assessment.

Repository Structure

This repository maintains three distinct branches representing different stages of the project lifecycle:

Branch Overview

develop main mvp-hackathon
Description Production-Ready PoC Pre-Competition Preparation Competition Demo Version
Scope • Cleaned and refactored codebase

• Professional code structure for PoC

• Optimized component architecture

• Comprehensive documentation

• Production deployment on Vercel 		• Initial preparation version (days before hackathon)

• Basic feature implementation

• Early integration with backend

• Foundational setup 		• Live competition demo version

• 24-hour hackathon build

• Full feature implementation under time pressure

• Direct integration with Amartha data

• Known technical debt from rapid development

Development Context: This branch represents the refined version developed post-hackathon, incorporating best practices, code cleanup, and architectural improvements. All rushed implementations from the competition have been refactored for maintainability and professional presentation.

Why not use mvp-hackathon as default?
The competition version was built under extreme time constraints (24 hours) and contains technical debt that needed addressing. This branch structure preserves the complete evolution from concept to polished implementation, serving as comprehensive technical documentation.

Branch Selection Guide

  • For PoC reference: develop (current branch)
  • For preparation study: main
  • For competition implementation: mvp-hackathon

Application Demo

Live Demo: https://app.social-collateral.id/

This is the production deployment of the refined PoC version, showcasing the complete Social Graph Engine with interactive visualization and multi-perspective risk analysis.

1. Interactive Social Graph Visualization

  • Real-time network rendering with Sigma.js
  • Intuitive zoom, pan, and node selection
  • Visual clustering based on geographic and social proximity
  • Color-coded nodes representing risk levels

2. Multi-Perspective Analysis Dashboard

When clicking a group node, the system displays comprehensive analysis through "Three AI Lenses":

  • Graph Analytics: Network centrality, clustering coefficient, community detection
  • NLP Insights: Sentiment analysis from field agent reports powered by Gemini AI
  • Computer Vision: Asset assessment from business/home photos via Google Vision API

3. Dynamic Filtering

Filter the network visualization by:

  • Kabupaten (District)
  • Desa (Village)
  • Group (Lending circles)

4. Trust Score Calculation

Algorithmic combination of:

  • Social graph structure analysis
  • Behavioral sentiment patterns
  • Economic asset indicators

License

Copyright © 2025 Tim Suksemustanice. All Rights Reserved.

This project is maintained as a technical portfolio and research documentation. The code, algorithms, and system architecture are proprietary intellectual property.

For commercial licensing, collaboration, or technical inquiries, please contact the team.

See LICENSE for full terms.