Features

GraphPolaris provides powerful customization and analysis options for adjacency matrices:

Visual Encoding

• Cell coloring: Use color intensity to represent connection strength, frequency, or other edge attributes.
• Color scales: Choose from sequential, diverging, or categorical color schemes based on your data type.
• Cell size variation: Optionally vary cell size to encode additional attributes.
• Row/column labels: Display node names, icons, or categorical indicators along axes.

Ordering Algorithms

• Alphabetical ordering: Simple baseline for finding specific nodes quickly.
• Degree-based ordering: Sort by connectivity to reveal hub patterns.
• Community detection: Automatically cluster related nodes to reveal dense blocks along the diagonal.
• Hierarchical clustering: Apply dendrogram-based ordering to show nested group structure.
• Custom ordering: Define your own sort based on node attributes like category, time, or importance.

Interactivity

• Cell hover: See detailed information about specific connections on hover.
• Row/column selection: Click labels to select all connections for a node.
• Region zoom: Zoom into specific areas of the matrix for detailed inspection.
• Brushing and linking: Selections in the matrix highlight corresponding elements in linked views.

Filtering and Aggregation

• Threshold filtering: Hide weak connections below a specified threshold.
• Attribute filtering: Show only connections matching specific criteria.
• Aggregated views: Group nodes by category to show summarized connection patterns between groups.

Comparison and Analysis

• Side-by-side matrices: Compare network structure across time periods or conditions.
• Difference matrices: Highlight what changed between two snapshots.
• Symmetry analysis: For undirected graphs, visualize only the upper or lower triangle.

Performance

• Virtualized rendering: Efficiently handle matrices with thousands of rows and columns.
• Progressive loading: Large matrices load incrementally for responsive interaction.

Example: An analyst comparing quarterly transaction patterns uses side-by-side matrices with difference highlighting to instantly spot new customer relationships and dropped connections.