Data

The Data panel is the entry point of the pipeline: it imports heterogeneous network data and builds the Blender graph object that every other panel operates on.

The panel is organized as a numbered, step-by-step flow. The exact steps shown depend on the selected data source.

Step 1 — Select data source

A row of toggles selects one of four ingestion pathways:

Source	Use it for
File	Flat files — GEXF (native graph), CSV/JSON edge lists and node tables, GraphML. The importer auto-detects native graph files.
Database	Live SQL connections (PostgreSQL, MySQL/MariaDB, SQLite; SQL Server optional) executing a `SELECT` query.
SuiteSparse	Matrices downloaded by identifier from the SuiteSparse Matrix Collection.
Repro	Replaying a declarative reproducible pipeline (see Reproducibility).

File source

Control
Description

Filepath
Path to the input file. If a native graph file (GEXF) is detected, a notice appears and structure mapping is automatic.

Delimiter
Column delimiter for CSV/tabular files (shown only for non-native files).

Load File
Parses the file and loads its columns so you can map graph structure.

Control	Description
Filepath	Path to the input file. If a native graph file (GEXF) is detected, a notice appears and structure mapping is automatic.
Delimiter	Column delimiter for CSV/tabular files (shown only for non-native files).
Load File	Parses the file and loads its columns so you can map graph structure.

Database source

Control	Description
Connection	Selects a saved database profile. The gear icon opens database preferences.
Test	Tests the selected connection.
Add New	Opens preferences to create a new connection profile.
SQL Query	The `SELECT` statement that returns the edge/node rows.
Load Columns	Executes the query metadata and loads available columns.
Preview	Previews query results.

Important

Only SELECT queries are allowed. SciGraphs never modifies your database.

SuiteSparse source

Control	Description
Matrix Identifier	`Group/Name` of the matrix (e.g. `Grund/bayer09`).
Graph Representation	`Symmetric` (standard adjacency via \(A + A^\top\), denser/rounder layouts) or `Bipartite` (rows and columns as separate node sets, preserving matrix structure with elongated layouts).
Giant component only	Keeps only the largest connected component via union-find, discarding isolated fragments.
Auto layout on import	Optionally computes a default layout immediately; pick the algorithm and scale. When coordinate data is present in the archive it is used directly and layout is skipped.
Download & Create Graph	Fetches the matrix and builds the graph in one step.
Browse Collection	Opens the online collection to find matrix identifiers.

Step 2 — Define graph structure

Shown for File and Database sources once data is loaded.

Control	Description
Source Nodes	Column providing the source endpoint of each edge.
Target Nodes	Column providing the target endpoint of each edge.
Directed Graph	Treats edges as directed (unlocks directed analysis and flow panels).
Remove Self-Loops	Drops edges whose source equals its target.
Weight Column	Optional numeric column used as edge weight.
Import as Attributes	A checklist of the remaining columns. Tick the ones to import as node/edge attributes; icons mark numeric, datetime, and text types. Use All / None to toggle the whole list.

Step 3 — Create graph object

Control	Description
Auto layout on import	Optionally lays out the graph immediately after creation (Default Layout + Scale). Skipped automatically in geospatial mode.
Create Graph	Builds the Blender graph object from the file or SQL results.

Importing node attributes

When a graph object already exists, an Import Node Attributes section lets you attach a vertex-only attribute table from a separate file.

Control	Description
Filepath	The attribute file.
Delimiter / Header	Parsing options for the file.
Import Node Attributes	Matches rows to nodes by name. Format: `node_name <sep> value(s)`; nodes missing from the file receive `NaN`.

Step 4 — Apply visualization

A shortcut to Setup Visual, which adds the Geometry Nodes rendering setup (instanced node spheres and edge cylinders). This is the same operation documented in the Visualization panel and is the bridge from raw topology to renderable geometry.

Subpanel — Geospatial Options

Note

When it appears: after a file is loaded or SQL columns are available. Enable it with the checkbox in the subpanel header.

Embeds the graph into real-world coordinates instead of using a force-directed layout.

Control	Description
Auto-Geocode Source/Target	Geocodes node names (countries/cities) to coordinates via Nominatim. The first run may be slow, then results are cached.
Latitude / Longitude column	Alternative to geocoding: use explicit coordinate columns.
Show Earth Globe	Adds a 3D globe; configure Globe Radius and Globe Quality (subdivisions).
Globe Theme	Texture theme: NASA Blue Marble (satellite), NASA VIIRS (night lights), Urban Dark, Topographic Shaded, Data Overlay (transparent), or none.
Texture Quality	Resolution of the downloaded globe texture.
PBR Settings	Water shine, water roughness, land roughness, surface relief (bump strength) for physically based globe materials.
Pre-download / Clear Cache	Manage the cached globe textures.
Material Style (theme = none)	Simple globe material, including a `World Map` mode with map detail and feature type.
Edge Style	Edge representation for geodesic arcs on the globe.

This subpanel drives the global-scale geospatial embedding used in the geospatial globe example.

Subpanel — Temporal Data

Note

When it appears: after a file is loaded or SQL columns are available. Enable it with the checkbox in the subpanel header.

Builds a sequence of graph snapshots from a time column and plays them back.

Control	Description
Time Column	Column holding timestamps.
Aggregation	How rows are grouped into snapshots; `Range` exposes From / To bounds.
Analyze Time Column	Inspects the time column and prepares snapshot creation.
Create Temporal Graphs	Builds the snapshot sequence (enabled after analysis).
Playback	Transport controls — first, previous, play, next, last — plus Refresh and Clear.

Typical workflow

Pick a data source in Step 1 and load it.
Map Source/Target columns and choose directedness in Step 2; tick any attribute columns to import.
Optionally enable Geospatial or Temporal options.
Create Graph in Step 3.
Setup Visual in Step 4 to make the graph renderable, then continue to Layout, Analysis, and Visualization.