The chart and the map, in lockstep.

Interactive Plotly charts rendered from feature attributes of layers on the map — with DataPlotty's signature bidirectional chart↔map selection linkage: lasso or box-select points in the chart and the corresponding features highlight on the map; select features on the map and the chart redraws. Six plot types, one fetch, OID-aligned selection. The QGIS DataPlotty plugin's capability, native to the desktop.

Scatter · Bar · BoxHistogram · Pie · ViolinBidirectional linkage Lasso → mapMap → chartGPKG config store

Bidirectional linkage

Select here, see it there

The linkage is the feature. A chart without the map connection is just a plot; a chart that drives and is driven by the map is an analysis tool.

Chart → map (Half A) — lasso/box-select or click points in the Plotly chart; the corresponding features are selected and highlighted on the map immediately. The OID travels with each trace as customdata, so selection is an O(1) lookup — never a re-query.
Map → chart (Half B) — select features on the map (or change the layer filter) and the chart redraws via Plotly.react (preserving zoom/pan/selection across the redraw).
Aggregating plots — histograms, bars, and pies select bins/categories, not points; the panel rebuilds the OID set client-side from the retained row data. No re-query on every click.

🔗

One fetch, OID-aligned

ChartData feature IDs travel as Plotly customdata. Selection events are O(1) OID lookups against the layer, never re-queries. A singular highlight handle (remove before re-set) keeps highlights from stacking. The layer's definitionExpression is captured-and-restored so "also filter layer" doesn't clobber the prior filter.

Six plot types

The charts analysts reach for

💧

Scatter

X/Y of two attributes — the classic correlation view. Click a point, see the feature.

📊

Bar

Categorical sums/counts. Select a bar, select every feature in that category on the map.

📦

Box

Distribution per category. Outliers are individually selectable.

📉

Histogram

Distribution of one attribute. Select a bin, select every feature in that range.

🥧

Pie

Proportion across categories. Click a slice, highlight its class on the map.

🎻

Violin

Distribution shape per category — the richer box-and-whisker.

Config that persists

Save a chart setup, re-open it later

Chart configs (plot type, layer, field bindings, layout) persist to a GeoPackage store — the same BookmarkStore pattern. Save the analysis setup, share it, or re-run it when the data updates.

💾

GPKG config store

Configs at <dataDir>/gpkg/chart_configs.gpkg, registered in gpkg_contents as attributes. Appears in the GeoPackage Manager + Connect-to-GIS catalog.

🔄

CRUD via REST

/api/charts/configs — list, create, update, delete. JSON properties/layout round-trip through JsonElement.

🔌

FeatureLayer + GraphicsLayer

The ChartSource abstraction backs both — one fetch path, OID-aligned, for served layers and ephemeral graphics alike.