Skip to main content

Box2Home - Internal Data Platform

View Frontend Source Code on GitHub →
View Backend Source Code on GitHub →

Overview​

What it is​

Internal platform for Box2Home delivery company providing developers with database access and debugging tools. Web-based interface for querying production data, viewing logs, and troubleshooting issues.

Why it exists​

Company relied on expensive third-party solution for database access and debugging. Needed cost-effective alternative with better performance and features tailored to team's workflow. Built as end-of-studies internship project.

Outcome​

Key Results
  • Replaced third-party tool - Eliminated costly subscription fees
  • Improved performance - Faster data queries and visualization
  • Better UX - Custom interface designed for team's specific needs
  • Audit logging - Complete tracking of database access and operations

Tech Stack​

Backend: NestJS (TypeScript), PostgreSQL
Frontend: React, Material UI
Deployment: Docker, AWS ECS
Authentication: JWT

Implementation Setup​

Backend (NestJS)​

Built modular backend following NestJS architecture:

Authentication & Authorization:

  • JWT-based authentication for developer access
  • Role-based access control (admin, developer, viewer)
  • Session management and token refresh

User Management:

  • CRUD operations for user accounts
  • Role assignment and permissions
  • User activity tracking

Data Access Modules: Created service/repository/controller pattern for each data entity:

  • Services: Business logic for data queries and transformations
  • Repositories: Database interaction layer with Prisma ORM
  • Controllers: REST API endpoints for frontend consumption

Audit & Operations Logs:

  • Log all database queries performed by users
  • Track data modifications and who made them
  • Store operation history with timestamps and user context
  • Searchable audit trail for compliance

Database:

  • PostgreSQL as primary data store
  • Read-only access to production database for queries
  • Separate logging database for audit trail

Frontend (React + Material UI)​

Built responsive web interface:

Features:

  • Data explorer: Browse and query database tables
  • Query builder: Visual interface for constructing queries
  • Results visualization: Tables, charts, and export options
  • User management: Admin interface for managing access
  • Audit logs viewer: Search and filter operation history
  • Authentication: Login flow with JWT token handling

Material UI Components:

  • Data tables with pagination, sorting, filtering
  • Forms for user management and query building
  • Navigation and layout components
  • Responsive design for desktop and tablet

Deployment​

Containerization:

  • Dockerfile for backend (NestJS)
  • Dockerfile for frontend (React production build)
  • Multi-stage builds for optimized image sizes

AWS ECS:

  • Backend service running on ECS Fargate
  • Frontend served from ECS with load balancer
  • Environment variables for configuration
  • CloudWatch logs for monitoring

Key Challenges & Solutions​

Challenge 1: Replacing Third-Party Tool Feature Parity​

Problem: Existing tool had features the team relied on. Needed to identify essential features and implement them without feature bloat while staying within internship timeframe.

Solution: Conducted user interviews with development team to prioritize features. Built MVP with core functionality (authentication, data querying, audit logs). Focused on most-used features rather than complete replication.

Result

Delivered platform with essential features in internship period. Team adopted new tool, eliminating subscription costs

Challenge 2: Ensuring Secure Database Access​

Problem: Platform provides direct access to production database. Needed to prevent accidental modifications and ensure only authorized developers can access sensitive data.

Solution: Implemented read-only database connections for data queries. Added role-based permissions with different access levels. Comprehensive audit logging of all queries. Required authentication for every request.

Result

Zero unauthorized access incidents. Complete audit trail of all database operations for security compliance

Challenge 3: Optimizing Query Performance​

Problem: Some database tables had millions of rows. Full table scans caused slow response times and poor user experience compared to previous tool.

Solution: Implemented pagination on backend with efficient SQL queries. Added database indexes for frequently queried columns. Frontend uses virtual scrolling for large result sets. Query result caching for repeated requests.

Result

Query response times improved significantly. Platform outperformed previous third-party solution in data visualization speed