Krakatoa
AI-assisted build (vibe coded) using Codex and Cursor. Built through iterative prompting and refinement; AI tools contributed to the implementation.
The problem
Volcanic history, terrain and monitoring data are difficult to understand when scattered across disconnected sources.
How I solved it
Built a persistent Three.js landscape across nine chapters. Integrated Open-Meteo, Darwin VAAC and Badan Geologi data with timestamp validation, separate observation/forecast layers and independent source-failure handling.
- Historical terrain transitions and public-data parsers
- Two APIs, automated tests and CI verification
The results
- interactive 3D atlas
- 9 chapters
- reviewed automated suite
- 13 tests
Delivered nine chapters, five historical states and six wind levels. Two APIs integrate three public-data sources; 13 automated tests and the reviewed test/build CI run passed.
Evaluation context. Educational visualization with illustrative terrain, not an operational warning system. CI success was observed in the profile review; deployment and scientific forecast accuracy were not verified.