🚀 Building a Fuel Route Optimization API with Django, Redis & OpenRouteService

Last night, I challenged myself to solve a very practical problem:

How do you find the cheapest fuel stops along a driving route between two cities?

At first, it sounded simple.

Until I opened the fuel station dataset and realized…

It had no coordinates. 😅

That’s when this turned from a simple API into a deep dive into:

• Routing
• Geocoding
• Spatial matching
• Caching
• Performance engineering

🧠 The Core Idea

Given:

Start Address → End Address as a string or list of coordinates

The system should:

1. Geocode both addresses using OpenRouteService
2. Fetch the driving route polyline
3. Build a route corridor using bounding boxes along the path
4. Match fuel stations from a CSV dataset that fall inside this corridor
5. Return the cheapest stations along the trip

No coordinates in the dataset? No problem. We make the map logic do the work.

⚡ The Performance Problem

Geocoding and routing APIs are expensive and slow.

So I introduced Redis caching:

• Addresses are geocoded once and cached forever
• Routes are cached
• Fuel stations are loaded into memory using Pandas

And here’s the fun part:

❌ No database
❌ No GIS server
✅ Pure spatial computation in memory

After the first request, response times drop to under 150ms.

🛠️ Stack Used

• Django + DRF
• Redis
• Pandas
• GeoPandas
• OpenRouteService API
• Spatial math (without PostGIS)

💡 Key Engineering Lessons

• Normalize messy data instead of bending your logic around it
• Cache aggressively when dealing with external APIs
• You don’t always need a database for spatial problems
• Bounding boxes + in-memory data can be incredibly fast

This project genuinely felt like building a mini map engine from scratch.