Data sources

A data source is where raw data first appears in a project. A data source stores the rows that come out of an upload or a plugin run, before any transformation.

Source types

• Excel — an .xlsx upload. The first row is treated as the header.
• CSV — a .csv upload with one row per line.
• Manual — an in-browser grid where you type values directly.
• Plugin — a feed from a plugin (SMHI, Google Sheets, Travsport, etc.). See Data Sources — Plugins.

Where sources sit in the pipeline

Data source → Mapping → Dataset → Template (via Production) → Renderer

The source is the raw truth. The mapping decides how its columns fill a schema. The dataset is the result. Templates and productions never read a source directly.

Refreshing

• Excel/CSV sources update when you re-upload.
• Manual sources update on every edit.
• Plugin sources refresh on their own schedule (or on demand from the source page).

After a refresh, every dataset derived from that source re-resolves automatically.

Why not skip to datasets?

A dataset can be manual too, bypassing a source. But for imported or live data, keeping the source separate makes it easy to:

• Re-import without rebuilding the dataset.
• Share one source between several datasets (filtered differently).
• See the exact raw values if the rendered graphic looks wrong.