← Back to home
Real estate2026

AirFloor

A floor-by-floor air-health score for any building in India, built to answer one real question before a flat purchase: which floor should you actually live on.

I replaced a floor-picking guess with a physics-informed air-health score for any building in India.

Design breakdown

How it was built, and why

AirFloor is a free, open-source web tool that scores every floor of a building in India for air quality, not just the address.

01

The flow

Pick a location, an India-only typeahead across area, PIN code, city, or district. Set the building's height with a segmented Floors, Meters, or Feet control and a draggable slider. The tool pulls live ground elevation for that spot plus live air-quality and meteorology data, then computes a 0 to 100 Floor Air-Health Index for every floor in the building. Results reveal as a curve, a bar view, and a tier table; tap or scrub any floor to pin it and see its full breakdown with a confidence band, not a single falsely-precise number. Below that, persona-specific advice for young children, elderly residents, asthma, and runners, and a two-tab buyer's checklist, questions to ask the developer and documents to request. Everything renders in English, Telugu, or Kannada, and the whole report exports to a clean multi-page PDF.

02

What it is

AirFloor is a free, open-source web tool that scores every floor of a building in India for air quality, not just the address. It started as a one-building calculator for a personal flat search in a 69-storey tower in Kengeri, Bangalore, then generalized into an all-India tool: type any place, set the building's height, and get an estimate of the healthiest floor to live on, built from live data instead of the usual assumption that higher just means better.

03

How it works

Built solo as a static site with no backend and no account: every computation runs client-side against live public data, so there is nothing stored and nothing to leak. The score itself is physics-informed rather than a flat higher-is-cleaner heuristic: it factors in inversion layers, wind penetration by height, and ventilation using live boundary-layer height and wind data, then runs a 48-scenario Monte Carlo pass so the result ships as a confidence band, not a single precise-looking number. The interface is a glass-morphism design with light and dark themes, fully localized into English, Telugu, and Kannada with natural rather than literal translations, built with real keyboard and screen-reader accessibility, and includes a real PDF export, an animated brand mark, and an India-only location lock so a search for a place outside the country the tool was built for is rejected rather than silently answered.

04

Why it matters

The trigger was concrete: choosing a floor in a 69-storey Kengeri tower with no way to check anything beyond higher floor, better view. AirFloor exists to replace that guess with an actual number, grounded in the same kind of physics used to model urban air pollution, applied literally at building height. Open-sourcing it, MIT license, public repo, a funding manifest, was the natural next step since the whole point was to hand a real answer to anyone else making the same expensive decision.

Components, live

The pieces, moving

AirFloor
Floor Air-Health Index
Every floor gets its own 0 to 100 index inside an 80% confidence band, not one falsely precise number. Scrub the tower or tap a floor to pin it.
0
index, floor 1 of 24
80% range 0 to 0
241
88best floor is 16. sweet spot 13 to 19, within 2 points of the peak.
56
PM2.5
35%
44
Fumes
20%
83
Vent
20%
82
Living
20%
92
O2
5%
floor 1 of 24|climbing the tower|best band 13 to 19