Polkadot already holds an open dataset of the planet's air: 19M+ readings from citizen sensors

People often say that one day data will live on chain. Here is a case where it already does, and where the main result of using a blockchain is not financial.

Since April 2026, home air sensors in Cyprus, on Koh Phangan in Thailand, on Réunion, across Europe and in 20+ countries have been sending their readings through Robonomics. Since 28 September, when Robonomics moved to Polkadot, the record lives on Polkadot.

What has piled up (counted on 13 November, from 1 April, when the current index starts):

  • 19M+ readings, 100M+ individual values: fine and coarse dust, temperature, humidity, pressure, street noise, CO₂ indoors
  • 100+ sensors in 20+ countries
  • almost half a million entries on chain

Earlier readings are on chain too and will be re-indexed, so these numbers are floors.

What the data has already shown

  • Koh Phangan: very clean background air, with about 1,200 short, sharp pollution episodes inside it, mostly in the evening and right next to a single sensor. That is daily household trash burning. Local activists now use it as evidence that sits on a public chain. Analysis: Koh Phangan, one year on: what a network built in two days actually found
  • Cyprus: 25+ sensors followed a Saharan dust storm hour by hour across the island, from Paphos to Nicosia and Famagusta. One official station shows one dot; two dozen neighbours show the front moving.

Why a chain

Each sensor signs its own readings. Batches go to IPFS, and the fingerprint of every batch is written to Robonomics on Polkadot. Change one number later and it no longer matches. The indexer is open source, so anyone can rebuild the whole archive from the chain without asking us.

Why it matters for Polkadot

This is a commons: a city, a university lab and a student get the same data on the same terms, under an open licence (ODbL), and nobody, including us, can close it. Polkadot here is infrastructure for a public good, not only for finance.

We would be glad to hear from researchers who want to use the data, and from anyone who wants to add a sensor to the map.

What is the actual point of the Treasury spending $200k on this? A standard centralized database could store 19M sensor readings for a few dollars a month.

Every day, Polkadot takes part in building a dataset of a kind that is rare: air measurements from people’s homes and streets in 20+ countries, signed by the sensors themselves and anchored on a public chain. It grows with every new reading, and nobody owns it.

That is what makes it a candidate for a public good, and it comes from how the data is stored, not from how much of it there is.

The same readings in a database on Microsoft’s cloud would be a private asset. Whoever pays the bill decides who gets access, on what terms, and for how long. They can edit a number or switch the service off, and users find out afterwards.

Here it works differently:

  • each sensor signs its own readings, and the fingerprint of every batch sits on Polkadot, so a changed value stops matching;
  • the indexer is open source, so anyone can rebuild the whole archive from the chain without asking us;
  • the licence is open (ODbL), so a city, a university, a student and a company get the same data on the same terms;
  • the readings live in IPFS and only small fingerprints go on chain, so keeping it costs very little.

The network keeps growing. If it reaches thousands of sensors around the world, which is a real possibility, Polkadot will hold a high-quality, verifiable air dataset open to anyone who needs air-quality data: researchers, cities, app builders, or just people who want to know what they breathe.

Anyone can copy it into their own database for a few dollars a month, and that’s welcome. The point is that every copy can be checked against the original, and the original belongs to no one.

The storage-cost comparison misses what the chain is buying: not bytes, but independent verification and continuity. A centralized database is cheaper until the operator changes terms, edits history, or disappears; signed readings anchored on Polkadot let anyone rebuild and check the archive. That said, if the proposal is $200k, the useful question is how much covers sensor support, indexing and maintenance versus on-chain fees — the latter should be tiny.