Back to Projects
HappyDeliveries | Real-Time Delivery Platform
A multi-sided delivery and personal shopping platform connecting customers, bike drivers, and local restaurants in Nelspruit. Built for a real client, funded by a local restaurant stakeholder.
Overview
1. The Problem:
Happy, a bike driver in Nelspruit, managed all orders through WhatsApp messages and phone calls. Customers would text their orders, Happy would manually relay them to drivers, and tracking was nonexistent. Payments were a mix of cash and informal card transfers, with no record of order history, driver performance, or popular menu items.
2. The Solution:
HappyDeliveries is a production multi-sided platform with three synchronized interfaces sharing one backend:
3. Customer App:
- Browse nearby restaurants filtered by proximity using PostgreSQL/PostGIS geospatial queries
- Add items to cart and place orders
- Choose payment method: PayFast (online), card-on-delivery, or cash-on-delivery
- Real-time order tracking: *order accepted → collected → in transit → delivered*
- QR-code-based delivery confirmation (5-minute expiry) to verify receipt
4. Driver App:
- Accept or decline incoming orders based on availability
- View customer details, delivery address, and order contents
- One-tap navigation handoff to Google Maps
- QR code scanner to confirm delivery completion
- Call or text customer directly from the app
5. Admin Dashboard:
- Manage driver onboarding and permissions
- Add and edit restaurant menus manually (partner restaurants are onboarded by the client)
- Set and update pricing per restaurant
- Analytics: most ordered menus, most active drivers, revenue collected per driver
6. Key Technical Decisions:
"Payment-on-acceptance logic:" Payment is only captured once a driver accepts the order. This eliminates the refund overhead that would occur if a customer paid online but no driver was available to fulfill the order.
"Real-time order tracking:" Socket.io handles live status updates between customer, driver, and admin interfaces without manual polling. PostgreSQL/PostGIS powers proximity-based restaurant filtering.
"QR verification:" When the driver arrives, the customer generates a QR code in their app (active for 5 minutes). The driver scans it to complete the order, preventing disputes over whether delivery occurred.
7. Business Context:
Funded in partnership with Coyotes Shisanyama, a local restaurant stakeholder in Nelspruit. The project was born from a real operational need observed in Nelspruit's informal delivery economy and is currently serving users in the area.
Happy, a bike driver in Nelspruit, managed all orders through WhatsApp messages and phone calls. Customers would text their orders, Happy would manually relay them to drivers, and tracking was nonexistent. Payments were a mix of cash and informal card transfers, with no record of order history, driver performance, or popular menu items.
2. The Solution:
HappyDeliveries is a production multi-sided platform with three synchronized interfaces sharing one backend:
3. Customer App:
- Browse nearby restaurants filtered by proximity using PostgreSQL/PostGIS geospatial queries
- Add items to cart and place orders
- Choose payment method: PayFast (online), card-on-delivery, or cash-on-delivery
- Real-time order tracking: *order accepted → collected → in transit → delivered*
- QR-code-based delivery confirmation (5-minute expiry) to verify receipt
4. Driver App:
- Accept or decline incoming orders based on availability
- View customer details, delivery address, and order contents
- One-tap navigation handoff to Google Maps
- QR code scanner to confirm delivery completion
- Call or text customer directly from the app
5. Admin Dashboard:
- Manage driver onboarding and permissions
- Add and edit restaurant menus manually (partner restaurants are onboarded by the client)
- Set and update pricing per restaurant
- Analytics: most ordered menus, most active drivers, revenue collected per driver
6. Key Technical Decisions:
"Payment-on-acceptance logic:" Payment is only captured once a driver accepts the order. This eliminates the refund overhead that would occur if a customer paid online but no driver was available to fulfill the order.
"Real-time order tracking:" Socket.io handles live status updates between customer, driver, and admin interfaces without manual polling. PostgreSQL/PostGIS powers proximity-based restaurant filtering.
"QR verification:" When the driver arrives, the customer generates a QR code in their app (active for 5 minutes). The driver scans it to complete the order, preventing disputes over whether delivery occurred.
7. Business Context:
Funded in partnership with Coyotes Shisanyama, a local restaurant stakeholder in Nelspruit. The project was born from a real operational need observed in Nelspruit's informal delivery economy and is currently serving users in the area.
tech-stack.json
// Technologies powering this project
const stack = [
"react-native",
"nodejs",
"express",
"postgresql",
"postgis",
"socket.io",
"payfast",
"real-time",
"delivery",
"mobile",
"rest-api"
];
// Interested in working together?
Let's build something.
I'm open to freelance, collaboration, and full-time opportunities.