How I built a feature-rich, zero telemetry, AI-powered user-friendly interface for Istanbul’s public transportation system using modern technologies
Introduction
I live in the bustling metropolis of Istanbul, where most of the 20 million people living there have to use public transportation — and a vast number rely on it daily. To plan your routine, coordinate transfers, or estimate your arrival time, having access to real-time transit data is not a luxury — it’s a necessity. This is where the iettnext project comes in: a privacy-first, modern application built specifically for the people of Istanbul.
In this article, I’ll walk you through the journey of building iettnext, from identifying the problems to implementing smart solutions. Whether you’re a developer, designer, or commuter interested in urban mobility, this story might resonate with you.
Table of Contents
- The Problem
- Tech Stack
- Architecture
- Key Features
- Development Challenges
- Performance Optimization
- Future Roadmap
- Conclusion
The Problem
Istanbul’s public transportation system, managed by IETT, is one of the most expansive and complex transit networks in the world. With thousands of buses, ferries, and metro lines moving millions of people every day, access to accurate and fast information is critical. However, the official tools — apps and websites — are often lacking:
- Outdated UI/UX: Difficult to navigate, not intuitive
- Performance Issues: High latency, unresponsive designs
- Lack of Features: No intelligent planning or real-time feedback
- Inaccessibility: Not optimized for all devices and users
Commuters deserve better — and so do the developers who want to build on this data.
The Beginning: Why We Started
I wasn’t just a bystander — I was also frustrated by the current tools. From tracking buses to estimating arrival times, the experience was fragmented, inaccurate, or just plain frustrating. My friends — a group of transit enthusiasts who loved tracking the IETT fleet manually — came to me with an idea: could I build something better?
With a background in cybersecurity and web development, I took the challenge seriously. I had lived in Istanbul long enough to know where the current systems failed. So I set out to build a tool that would be:
- Transparent (open source)
- Private (no telemetry)
- Modern (fast, mobile-first, AI-enhanced)
It all started on May 1st, with the goal to ship something functional fast — but not rushed.
Designing the Solution
To meet the needs of a dynamic and growing user base, I approached iettnext like any good system: modular, scalable, and user-focused. Some design principles I stuck to:
- Zero telemetry: I wanted users to feel safe. No tracking, no hidden logging, no marketing scripts.
- Fast-first: Transit apps should be lightweight and instantly responsive.
- Mobile-first UI: Istanbulites are on the go — so was the interface.
- Expandable by design: I didn’t just want a bus tracker. I wanted a foundation for a full transportation assistant.
Tech Stack
I opted for a modern stack that could give me rapid development without sacrificing stability:
- Backend: Deno — A secure, fast, TypeScript-native runtime
- Database: PostgreSQL — Robust, relational, and open source
- Frontend: Plain HTML/CSS with Tailwind — Clean, readable, and lightweight
- Mobile App: Built using Expo (React Native), open-sourced for transparency
The result was a snappy app with clean separation of concerns, where each module handled its role efficiently.
Key Features
The project quickly evolved into a powerful tool with features including:
- Real-Time Bus Tracking
- AI Chat Assistant (to answer transit queries naturally)
- Vehicle Detail Pages (with model, speed, garage info, etc.)
- Route Lookup & Timetables
- Historical Task Data for Buses
- Zero-telemetry privacy design
- Offline-capable Android App (with downloadable APK)
These features weren’t just implemented — they were refined based on user feedback.
Development Challenges
No journey worth telling is without obstacles. Here are some of the biggest challenges:
- Unreliable APIs: The official endpoints were undocumented, sometimes inconsistent
- Time constraints: The project had a tight timeline (initial release aimed in under a month, single developer working on the system)
- Concurrency issues: Handling real-time data with concurrent users required caching, retries, and rate awareness
- Mobile UX: Building something truly intuitive on smaller screens took iteration
Still, the hardest challenge was resisting feature creep while maintaining quality.
Performance Optimization
From day one, performance mattered. To keep the app light and responsive:
- Server-side caching was used to reduce API hits
- Modular fetch logic ensured only necessary data was queried
- Database indexing improved read times
- Lightweight UI avoided bloated libraries
The end result: sub-second response times for most queries, even under load.
Future Roadmap
iettnext is still growing, and here’s where it’s headed:
- Metro and ferry integration
- Dark mode refinements and accessibility boosts
- User accounts for personal tracking (opt-in, privacy-preserving)
- More AI features: smarter recommendations, speech-to-text
- Progressive Web App (PWA) deployment
Conclusion
What started as a simple tool for a few friends has grown into a modern, useful application for thousands of Istanbulites. It’s fast, open, and powered by real need.
My hope is that iettnext continues to inspire confidence in the city’s transportation — one ride, one request, one tap at a time.
If you’re building something similar, or just want to make your city better, I’d love to hear from you. Open source, after all, is better together.
Thank you for reading. Let’s keep building tools that matter.