Climate Quantum

Climate Quantum Inc. is not a real company. It is the fictional enterprise I use as the running case study across two O’Reilly books — Building Data Products and Implementing Data Mesh. Every data contract, data product, and marketplace example in those pages is built on it.

Why a climate company?

Because climate data is the hardest data problem I could hand a reader. It is not static — it shifts constantly under both human activity and natural phenomena. It arrives from thousands of sources, from agencies like NASA, NOAA, ESA, and ECMWF down to individual weather stations, each with its own format, its own scale, and its own licensing terms. It spans meteorology, oceanography, glaciology, and environmental science, so integrating it is a genuinely multidisciplinary problem.

And the stakes are unusually high. Regulation keeps expanding, with Scope 1, Scope 2, and the notoriously awkward Scope 3 emissions all demanding precise reporting. Worse, the assumption underneath most forecasting — that the past predicts the future — is failing. Insurers priced California wildfire risk and Florida hurricane risk on historical patterns, then left both markets when reality outran the history. When decisions of that magnitude rest on your data, accuracy stops being a technical concern and becomes a moral one.

A centralized data platform buckles under that load. That is precisely why it makes such a good teaching case for a decentralized, domain-oriented architecture.

The mission

Climate Quantum Inc. exists to make climate data easy to find, consume, share, and trust. Those four verbs drive every design decision in the books:

• Find — vital datasets are scattered across countless sources. A curated catalog of data products, backed by discovery APIs, turns the hunt into a search.
• Consume — inconsistent formats and structures block real use. Standardized access methods give researchers, policymakers, and businesses one way in.
• Share — collaboration dies without protocols. Explicit data contracts spell out how data is both consumed and published.
• Trust — governance certifications verify origin, lineage, and quality, so a consumer knows what they are relying on.

How it is built

The architecture has a few deliberate quirks, and they are where most of the teaching happens:

• A global climate data mesh supporting hundreds of data products, each owned by a largely autonomous team responsible for its own slice of climate data.
• A Climate Data Marketplace that works like DNS for climate data — a single directory that surfaces products from across the whole mesh.
• Data products that reference rather than copy. The volumes are too large and change too fast to duplicate, so products link to data where it already lives. This one decision avoids a cascade of lineage and governance problems.
• A uniform geospatial index — Uber’s H3 hexagonal schema — so every data point, whatever its type, aligns to one consistent view of the planet.
• Publishing in five minutes or less. Most platforms optimize for consumers; Climate Quantum optimizes for publishers, on the theory that more publishers eventually means more consumers.

---

The two books

Climate Quantum runs through both. Read either on its own, or both in order.

No O’Reilly subscription? You can get one free for a month, and 20% off, with code JEAN20.

Building Data Products

Increase Trust and Unleash Data-Driven Innovation — forthcoming from O’Reilly Media. An early edition is available now.

Implementing Data Mesh

Design, Build, and Implement Data Contracts, Data Products, and Data Mesh — O’Reilly Media, September 2024. Written with Eric Broda, foreword by Scott Hirleman.

---

See all my books, or come hear me speak.