Daniel Shavit: show me the decision behind this street

The money was in the spreadsheet. Some of the names weren't.

Daniel Shavit began his Futureproof presentation with a familiar kind of civic frustration. A political argument over dinner in Paris led him to look for the public data that might settle it. He found records that raised more questions: payments without useful context, separate datasets that didn't connect and information he had to chase through other documents.

Putting a CSV online can satisfy the act of publishing while leaving a citizen with a second research project just to understand the first one.

Daniel's work asks what would happen if we started with the things people already care about. The park nearby. The theatre. A contract for work down the street. Who made the decision? What did they approve? Where can somebody read it?

That question became Tale of a City, his project through the AI Builders Fellowship with Internet Archive Canada and BC + AI.

Daniel described rebuilding connections between public spending and the places it concerned. In his Paris example, a theatre becomes an entrance to a history of grants, works and council deliberations. Someone doesn't have to know the document's filing vocabulary before they can begin.

A map can make a relationship easier to see. It can also make a claim look more settled than it is. The important move here is keeping a route from the place back to the decision and from the decision back to its source.

Our October feature on the two fellows' projects described that geographic route alongside Jacob Ortiz's approach to comparing government publications. They are different entrances into the public record. Both become more useful when a reader can leave the summary and inspect what it rests on.

At Futureproof, Daniel brought the work to Vancouver. He described getting access to years of council minutes through the Archive collaboration, then showed a film combining a city view with the records connected to it.

The film gives the place visual presence. Its more consequential move is bringing documents and archive material into the view of a particular street. A city stops being only a surface to look at. You can begin asking how decisions accumulated there.

There is a distinction worth keeping beside those images. A compelling historical view is still a model with choices about dates, buildings and missing information. On the current Vancouver project page, the method explains that the dated buildings are buildings still standing, rather than a complete reconstruction of every earlier city. It also names gaps in council records and in the underlying map data.

That is useful information to put where a visitor can find it. A year slider can look wonderfully authoritative. It needs an explanation of what changes when you move it and what the map cannot show.

In my first-year account of Futureproof, the fellows were part of what made opening night feel concrete: people had made things we could look at and question. Daniel's film creates that invitation. The source links and method have to carry it further.

After the Vancouver film, Daniel showed a San Francisco example connecting older records to a current civic agenda. He described tracing who was involved, where money went and what earlier decisions concerned the issue now coming forward.

This is where a civic archive can change the work required to participate. A journalist investigating an issue, a resident preparing a question and a public employee checking a decision all need context. If each person has to rebuild the history alone, the people with time and institutional knowledge begin with an enormous advantage.

Daniel's proposal is to make more of that history available at the point where somebody can use it. He also showed a question-and-answer interface with source references alongside the response.

On stage, he argued that organizing the underlying dataset could make it usable by different models, including ones people run themselves. That is a design ambition worth examining. Preserving a usable dataset, its sources and a way to query it gives future builders something they can work with when today's preferred model changes.

He said he wanted to keep the decisions updated and make extracted information available beside the original records. The public Tale of a City repository offers a place to inspect the Vancouver data work behind the presentation. The upkeep matters as much as the first demonstration. A civic tool becomes less useful if the city keeps making decisions while its map stands still.

Pick a place you know well. Find the record behind one decision that shaped it. Then follow the next decision still being made. A public archive earns a different kind of value when it helps somebody arrive at city hall with a better question.