Zhanik on Stardance

Ship

I built Aetheris, an environmental operating system for Kazakhstan that visualizes live data like AQI and PM2.5 on an interactive map and explains it using a custom AI analyst. The biggest challenge was rendering massive volumes of sensor data without UI lag and engineering the AI prompts to ground answers in local context, like Almaty's winter temperature inversions, rather than outputting generic facts. I am most proud of this reasoning layer because it transforms raw, disconnected environmental readings into plain-language, actionable insights. To test the project, open the "Atlas" on the live site to explore the different environmental layers, then head to the AI Assistant and ask something specific like, "What is the risk outlook for Almaty?" to see how it interprets the live map data.

1 devlog
4h

Try project → See source code →

Open comments for this post

@Zhanik on Aetheris · 1 day ago

3h 33m 30s logged

Day 6 Update: Live AQI Map & AI Analyst for Aetheris 🌍🇰🇿

Today I spent my session building out the core infrastructure for Aetheris. My main goal was to bring the Kazakhstan Atlas to life and connect it with our AI reasoning engine.

What I built today:

Set up live data fetching for AQI and PM2.5 across 28 cities in Kazakhstan.
Implemented the logic for the Environmental AI layer. The assistant now processes real-time sensor data and explains the context (e.g., the heavy smog caused by winter temperature inversions in Almaty).

Challenges faced:
Processing such a massive volume of environmental data points caused severe UI lag during the initial map render. I resolved this by optimizing the data arrays and implementing server-side caching.

Next steps:
In the next session, I will focus on the Community system so citizen scientists can start logging local pollution events and earning contribution points.