Fırat Group · Türkiye

Baran Fırat Şahin

I build and operate live systems on my own — from the first line of code to production. The screen above is one of them. This page explains what it is, how it was built, and where every number on it comes from, in my own words.

What this system is

The World Terminal is a single-surface operations screen: one map, one clock, one continuous stream of real events. It exists because I wanted to see the state of the world the way an operations desk sees it — not as a feed of articles, but as positions, magnitudes and timestamps on a map you can zoom into until you reach street level.

It is written by hand. No framework, no page builder, no template, no component library. The interface is plain HTML, CSS and JavaScript; the backend is a single edge worker that fetches, normalises, caches and serves data close to the visitor. I made that choice deliberately: a framework would have been faster to start and slower to understand, and I wanted to understand every layer of what I was running.

The one design rule: real data, honest failure

Every value on the screen resolves to a live source. Nothing is simulated, padded, or invented to make the interface look busy. When a source stops responding, the terminal does not substitute a plausible number — the panel switches to SIGNAL LOST and stays there until the source recovers.

This sounds like a small thing. It is the hardest constraint in the whole project, and it is the reason several features were cut. A dashboard that fills gaps with invented values is a dashboard that lies quietly, and a screen that lies quietly is worse than an empty screen. The visitor counter is a real counter. The event log is a real log. If a number is on this page, something measured it.

How it is engineered

• Edge-first delivery. The interface and the data proxy run at the network edge, so first paint and first data arrive together rather than in sequence.
• Multi-source resilience. Each layer has more than one upstream provider. When the primary fails, the worker falls back; when all of them fail, the layer reports its own failure instead of hiding it.
• Cache discipline. Upstream providers are called on a fixed schedule and cached at the edge, so a thousand visitors do not become a thousand upstream requests. Being a good client of a public source is part of the design, not an afterthought.
• Geospatial rendering. Positions, orbital tracks, event radii and city markers are drawn on a deep-zoom map surface with layered tiles, tuned so the screen stays readable at world scale and still resolves at street scale.
• Operated solo. Deployment, monitoring, incident response and every fix on this site are mine. There is no team behind it and no agency.

Where the data comes from, and what is mine

The underlying measurements — seismic events, natural hazards, orbital and air traffic positions, atmospheric and space-weather indicators, market and network indicators, and news headlines — belong to public scientific, governmental and publisher sources. I do not own that data and I do not claim to. Every record carries its source inside the application: the panel names it, the popup names it, and each news headline links directly to its original publisher.

What is mine is the system around the data: the interface, the normalisation and fallback logic, the failure semantics, the geospatial correlation between events and places, the caching layer, and the judgement about what belongs on a single screen and what does not. Headlines are shown as headlines and are never republished as full text. The purpose of this site is to demonstrate the engineering, not to reproduce someone else's content.

Who I am

I am an independent builder from Türkiye, working across offensive security, fullstack and edge engineering, applied AI, and market infrastructure. I run IT operations at Fırat Group / BEQW Systems. Security work is authorised assessment work only — penetration testing, vulnerability assessment, hardening and security automation — and its details stay private for operational reasons. On the AI side my interest is systems that run fully offline on local hardware, where no data leaves the machine. On the market side I build systematic infrastructure with hard risk controls; strategies and parameters are private.

Most of what I know is self-taught, built over years of shipping things that had to actually run. I hold degrees in International Marketing (Selçuk University) and International Trade & Logistics (Anadolu University).

Contact

LinkedIn: Baran Fırat Şahin · GitHub: github.com/baranfiratsahin · Detailed profile →