Skip to content

Latest commit

 

History

History
92 lines (63 loc) · 3.62 KB

README.md

File metadata and controls

92 lines (63 loc) · 3.62 KB

Collections Queue Dashboard (PoC)

Status License

A full-stack "Vertical Slice" Proof-of-Concept for a high-volume financial collections system.

This project demonstrates a prioritized work queue focusing on SQL performance (SARGable queries), data integrity, and Optimistic UI patterns. It was built to solve the specific problem of "delinquency interception" by reducing agent cognitive load and dashboard latency.

📸 The Vertical Slice

The UI (React 19) The API (Swagger/OpenAPI)
React Dashboard API Swagger
Optimistic updates with "Undo" safety net Minimal API with Dapper & SQL Server

🏗 Tech Stack & Architecture

  • Database: MS SQL Server 2022 (Dockerized)
  • API: .NET 9 Web API (Minimal API pattern)
  • ORM: Dapper (Micro-ORM for raw SQL performance)
  • Frontend: React 19 + TypeScript (Vite)
  • State: React Hooks + Optimistic UI updates

🚀 Key Features & Engineering Decisions

1. Performance-First SQL (SARGable Queries)

The core logic resides in sp_GetPriorityQueue.

  • The Problem: Typical DATEDIFF functions in WHERE clauses force full table scans.
  • The Fix: I calculate the @CutoffDate variable before the lookup.
  • The Result: The SQL engine performs an Index Seek, allowing the system to scale to millions of records without latency.

2. Vertical Slice Architecture

Instead of generic "Repository" and "Service" layers that create horizontal coupling, this app is built in Vertical Slices:

  • Read Slice: GET /api/queue → Dapper Query → SQL Stored Proc
  • Write Slice: POST /api/queue/{id}/log → Dapper Command → Parametrized SQL

3. Optimistic UI & Error Recovery

  • Instant Feedback: The moment an agent logs a call, the item vanishes from the queue (client-side filter).
  • Safety Net: An "Undo" Toast notification allows recovery from accidental clicks without a page reload.

📂 Project Structure

CollectionsQueue-PoC/
├── api/                  # .NET 9 Backend
│   ├── CollectionsApi/   # Web API Project
│   └── sql/              # Database Setup Scripts (Schema & Seeds)
├── ui/                   # React 19 Frontend
│   └── collections-ui/   # Vite Project
└── docs/                 # Screenshots and documentation

🛠 How to Run Locally

Prerequisites

  • Docker Desktop
  • .NET 9 SDK
  • Node.js (LTS)

1. Database Setup

First, spin up the SQL Server container. (Note: You can change the password below, but ensure you update appsettings.json in the API project to match.)

docker run -e "ACCEPT_EULA=Y" -e "MSSQL_SA_PASSWORD=YOUR_PASSWORD_HERE" -p 1433:1433 --name sql_server_dev -d mcr.microsoft.com/mssql/server:2022-latest

Once running, connect via Azure Data Studio (or SSMS) and execute the setup.sql script located in api/sql/setup.sql.

2. API Setup

Navigate to the API folder and start the server.

cd api/CollectionsApi
dotnet run

3. Frontend Setup

In a new terminal, navigate to the UI folder.

cd ui/collections-ui
npm install
npm run dev

This project is a portfolio demonstration of full-stack architectural patterns.