Appearance
Exporting data
kepler's Export Data writes the map's datasets to a file, in the browser, from whatever the panel holds at that moment. It is in the export menu behind the download icon at the top of kepler's side panel.
Formats
| Data type | What you get |
|---|---|
csv | a CSV |
geojson | a GeoJSON feature collection |
parquet | a GeoParquet file |
shp | a Shapefile, zipped |
gpkg | a GeoPackage (SQLite), in EPSG:4326 |
kml | a KML |
kmz | a zipped KML |
The dialog also asks for the Dataset — all of them, or one — and whether to write the Filtered rows, the ones kepler's filters leave on the map, or the Unfiltered ones, every row the query returned. The file is named after the dataset: kepler.gl_Query-A.gpkg for one called Query A.
Nothing leaves the browser
The file is written where the map is drawn. GeoPackage and GeoParquet need a writer compiled to WebAssembly, sql.js and parquet-wasm, which kepler on its own fetches from cdn.jsdelivr.net and unpkg.com at the moment of exporting; the plugin ships both and serves them from Grafana, under public/plugins/ubica-keplergl-panel/loaders/, so an export asks no other host for anything and a hardened Grafana needs nothing added to its Content-Security-Policy for it. See Differences from stock kepler.gl.
Who can export
Anyone who can open kepler's side panel, viewers included: the export is kepler's, and it knows nothing of Grafana's roles. What it writes is only what the query already sent to that browser, which Grafana's own Inspect → Data offers too. So the way to keep data from a dashboard's readers is the query: leave out the rows and columns the map does not need.
Upstream: Save and export
Written against kepler.gl 3.3.0-alpha.17. See Upstream documentation.