What’s Next for Histomap Explorer

One of the most interesting discoveries during the development of Histomap Explorer didn’t come from music history at all. It came from petroleum.

As the project grew beyond music, I started pointing it at entirely different kinds of historical subjects. One of the first was petroleum and petrochemicals — a visual history of how crude oil went from a source of lamp fuel to one of the foundational building blocks of modern life. The idea seemed straightforward. Histomap Explorer already knew how to visualize the rise and fall of musical genres, artists, and influential works, so why not point the same machinery at petroleum-derived products, industrial processes, and supply chains?

The first draft came together fast. Within a short time the system produced a working map tracing the progression from extraction and refining out into transportation fuels, plastics, synthetic fibers, medical materials, electronics, and modern energy-transition technologies. The result was fascinating — and it was also difficult to validate.

With music, I can usually look at a map and judge intuitively whether it feels right. If a Progressive Rock ribbon suddenly dominates the 1990s, I know enough to question it. Petroleum was different. The map looked plausible and the story it told was coherent, but I’m not a petroleum historian, a refinery engineer, a petrochemical economist, or an industrial chemist — and most future users of Histomap Explorer won’t be either.

That’s the catch. The very reason someone requests a new domain is usually that they want to learn about it. They’re curious, not expert. And that uncovered a hidden opportunity in the way I’d been generating domains — one I can frame as a single question:

The fashionable answer today would be to reach for an AI expert: spin up a specialized agent, steep it in the domain, and let it generate the map. But that just swaps one problem for another. An AI that sounds authoritative isn’t necessarily correct.

For domains that touch economics, science, politics, technology, medicine, or industrial history, confidence is not evidence — and the more persuasive the output looks, the more it matters to know where its conclusions actually came from. A convincing wrong answer is worse than an obviously rough one, because nobody thinks to check it. Histomap Explorer needed something different.

Instead of building an “expert agent,” I started exploring the idea of a Domain Steward. The distinction is the whole point: a steward isn’t treated as an authority. It’s treated as a researcher. Its job is to gather information, identify sources, propose a historical structure, document its assumptions, and explain why the map looks the way it does.

Every new domain would begin with a steward working through the same set of questions: What is the scope of the domain? What are the major historical movements? Which entities, people, organizations, inventions, or technologies matter most? What evidence supports the rise and decline of each movement? And what assumptions were required to turn all of that into a visualization?

Rather than hiding its reasoning, the steward exposes it. Every ribbon, every milestone, every historical claim carries a documented rationale you can trace back.

As the idea developed, a second role emerged. If the steward’s job is to build a case, someone has to be responsible for challenging it. That became the Adversarial Reviewer — a role whose entire purpose is to ask the uncomfortable questions. Are the sources reliable? Are important perspectives missing? Are categories being double-counted? Does the visualization imply more precision than the evidence supports? Are politically or economically charged topics being handled fairly? Is the map teaching history, or quietly arguing a thesis?

In the petroleum experiment, this stage earned its keep. The review surfaced things I’d never have caught on my own — for example, that petroleum and natural gas were being blended together too casually, that petrochemical feedstocks were bleeding into downstream industries like plastics and synthetic fibers, that some ribbon widths looked reasonable but had no source-backed justification, and that economic value and material dependency were quietly being treated as the same thing. None of it invalidated the map. All of it made the map better.

What emerged was a different vision for how Histomap Explorer could grow. Instead of generating a new domain in one shot, future domains would move through a structured pipeline — and the output wouldn’t just be a map. It would be a map with provenance, where a visitor can explore not only the visualization but the research process that produced it.

What makes this work as a product is that the person asking only ever touches the two ends. They make a request; they get a trustworthy map. Everything in between — the steward’s research, the reviewer’s challenges, the back-and-forth that resolves them — runs behind the curtain, there to inspect if they ever want to, invisible if they don’t.

I think of that pipeline in four phases:

The exciting part is what this does to the whole project. It turns Histomap Explorer from a visualization tool into something larger.

Picture a visitor asking for a domain that doesn’t exist yet — the history of vaccines, the evolution of semiconductors, the development of renewable energy, the history of space exploration, the rise of global trade. Rather than waiting for a human expert to hand-build each one, Histomap Explorer could materialize a temporary Domain Steward to research the topic, assemble the evidence, propose a historical model, and document its assumptions. A reviewer would then challenge those conclusions before any of it reached the public collection.

The result is a genuine collaboration between human curiosity and AI-assisted research. The curiosity sets the direction; the research, the challenge, and the documentation make the result something you can actually stand behind.