Documentation → How-to guides

Build a timeline

A timeline is a dataset whose features carry a validity interval — the year each was built and, optionally, the year it was retired. A time slider then scrubs the map through the build-out: features appear the year they arrive and drop out the year they go. It is the natural shape for a network that grows — stations, cables, sites, permits — where each feature is added rather than moved.

How a timeline is modelled

Every feature keeps one fixed geometry and two number fields:

Built year — the year the feature becomes live (e.g. 2028).
Decommissioned — the year it retires. Leave it empty for a feature that is still live.

A feature is shown at year Y when it was built on or before Y and has not yet been decommissioned. A network doesn't grow by reshaping a feature — it grows by adding more features, each with its own built year. A feature with no built year is a baseline: it is treated as always there, so it shows at every year. Years are whole numbers; months and days are not modelled.

A timeline is a normal dataset. It lives beside your uploads, can be styled and put on any map, and feeds a SQL model like any other. What makes it a timeline is the pair of year fields and the slider they switch on — nothing else changes about how it is stored, shared or queried.

Create one from scratch

Draw a timeline directly on a map when you are authoring the features yourself.

Open a map in the editor and choose Draw features to start an editing session.
Choose New Timeline (the clock, below New Layer). A fresh, empty timeline is added to the map and opened for drawing.
A year slider appears along the bottom of the map. Set it to the year you are drawing — it reads New at year while a timeline is open, because every feature you draw is stamped with that year as its built year.
Draw your features. Move the slider forward and keep drawing to lay down later years; each feature carries the year it was drawn at.
Choose Save. The timeline is tiled in place and the slider now scrubs through what you have built.

The year on the slider is only a default for new features — it never rewrites a feature you already placed. To change a year after the fact, edit the feature (below).

Fork one from an existing dataset

If you already have the geometry — a layer of sites, say, with a build-year column — turn it into an editable timeline instead of redrawing it.

Open the dataset and choose Fork as Timeline.
Choose how each feature gets its built year:
- From a field — map an existing numeric column (a year) onto the built year.
- One year for all features — treat the whole dataset as built in a single year.
- No build year — every feature starts as a baseline; set the years yourself in the editor.
Choose Fork as Timeline. A new, editable timeline is created — the original is left untouched — and begins tiling.

Forking is for small sources. A fork copies the features into an editable object, so it is capped at 2000 features and 2 MiB. A larger layer is a job for a full dataset, not a hand-edited timeline, and the fork is refused with that nudge. Any attribution and licence obligations on the source travel to the fork automatically — a forked timeline is a derived dataset and carries the same credits.

Edit the years on a feature

Select a feature while editing. Above its ordinary attributes you get the two timeline fields:

Built year — type a year, or leave it blank to make the feature a baseline (the field shows baseline when empty).
Decommissioned — choose Decommission at year to retire the feature at the slider's current year, or type a year directly. Clear it to mark the feature still live again.

To edit many features at once, switch the attribute editor to Table. A timeline's table always leads with both year columns — built and decommissioned — so you can fill in a retirement year inline even while every feature is still live. As always, choose Save to re-tile.

What the slider shows

Dragging the slider re-draws the map instantly — no re-processing, no waiting. At each year a feature falls into one of four states:

State	When	On the map
New this year	built exactly this year	Highlighted, so this year's additions stand out from what was already there.
Live	built, not yet retired	Your normal styling for the layer.
Retired	decommissioned this year or earlier	Ghosted, so you can still see where it was while editing.
Future	built after this year	Hidden — it hasn't arrived yet.

Compare two timelines

Once a timeline is saved it drops straight into Compare — the side-by-side swipe view launched from the maps list. Pick two timelines, drag the divider to hold them against each other, and scrub the shared year to watch two build-outs diverge. In the comparison, retired features simply drop out at their decommission year rather than being ghosted.

Troubleshooting

I drew a feature but it is dated years in the future: Check the year on the slider before drawing — it is the built year every new feature gets. Set it to the year you mean first, then draw. To fix one already placed, select it and change its Built year.
A feature never disappears when I scrub forward: It has no decommission year, so it stays live for good. Give it a Decommissioned year (or use Decommission at year) and Save.
Every feature shows at every year, whatever the slider says: Those features have no built year, so they are baselines — always visible by design. Set a built year on each one you want to gate by time.
Fork as Timeline is refused, or the button is missing: Forking is offered only on your own dataset once it is Ready, and only for small sources (up to 2000 features / 2 MiB). A larger layer, a still-processing one, or a borrowed reference layer cannot be forked — bring the large ones in as a full dataset instead.
The years won't save: Built and decommissioned must be whole years, and a feature cannot be decommissioned before it was built. Fix the offending value — the save reports which — and try again.