Tree Plot

Use me when the relationships between things matter as much as the things themselves. I draw your hierarchy as a clean, connected tree — parent above child, edges showing who reports to whom or what belongs to what. Org charts, taxonomies, family trees, decision paths — if structure is the story, I tell it clearly.

Overview

A tree plot renders hierarchical data using the Reingold-Tilford layout algorithm, which produces evenly spaced, non-overlapping node arrangements that make parent-child relationships immediately legible. Nodes are connected by edges, and the tree can grow top-down, bottom-up, left-right, or right-left. Unlike treemaps and sunburst charts that encode proportion through area or angle, a tree plot encodes structure — who is connected to whom and at what depth.

Best used for:

• Organizational charts (CEO → VP → Director → Manager)
• Biological or scientific taxonomies (Kingdom → Phylum → Class)
• Decision tree visualization from a trained model
• File system or folder hierarchy navigation
• Family trees and genealogy
• Network dependency graphs with a clear root

Common Use Cases

Organizational & Team Structure

• Company org charts — who reports to whom at every level
• Team rosters grouped by department and role
• Project accountability maps (project → workstream → owner)

Science & Classification

• Biological taxonomy (Domain → Kingdom → Phylum → Class → Order)
• Chemical compound family trees
• Medical condition classification hierarchies

Software & Systems

• File system directory trees
• Software module and package dependency graphs
• Microservice ownership hierarchies
• CI/CD pipeline stage breakdown

Data Science & ML

• Decision tree model visualization after training
• Feature split paths in a random forest
• Clustering dendrogram exploration

Knowledge & Content

• Topic taxonomy for a content library
• Legal document hierarchy (Act → Chapter → Section → Clause)
• FAQ category trees

Options

Path / Hierarchy

Required - Select two or more categorical or text columns in order from root to leaf.

The order of your selections defines the tree depth:

• First column = root level (e.g., Continent)
• Second column = level 2 (e.g., Country)
• Third column = level 3 (e.g., City)

Each unique combination of values across the selected columns becomes a node in the tree. Selecting 2 columns produces a 2-level tree; 3–4 columns are recommended for most use cases. Beyond 4 levels the tree can become dense — consider filtering or splitting into subtrees.

Value / Size

Optional - Size nodes by an aggregated numeric metric.

When configured, each node's visual size (marker diameter) scales with the aggregated value of the selected column. Useful for showing magnitude alongside structure — for example, sizing department nodes by headcount.

Column

Select a numerical column to aggregate (e.g., Headcount, Revenue, Score).

Aggregation

• Sum - Total value rolled up through the hierarchy
• Average - Mean value at each node
• Count - Number of records under each node
• Min / Max - Range extremes at each node

Settings

Hide Empty Values

Default: Off — When enabled, nodes whose value field is null or empty are excluded from the tree entirely. Useful when your hierarchy has structural gaps (e.g., some regions have no cities in the dataset).

Advanced Options

Color By

Optional - Color nodes by a categorical, text, or numerical column.

For categorical columns, each unique value gets a distinct color — handy for highlighting department type or status. For numerical columns, a continuous color scale is applied so you can read both structure and intensity simultaneously.

Hide Labels

Default: Off — Suppresses text labels on all nodes. Hover tooltips still show node names and values. Useful when the tree is very dense and labels overlap.

Orientation

Controls the direction the tree grows:

• Top to Bottom (default) — Root at top, leaves at bottom. Best for org charts and standard hierarchies.
• Bottom to Top — Root at bottom, leaves at top. Occasionally used for evolutionary trees.
• Left to Right — Root on left, leaves on right. Best when node labels are long and need horizontal space.
• Right to Left — Root on right, leaves on left. Mirror of left-to-right.

Node Size

Sets the default diameter for all nodes when no Value/Size metric is configured:

• Small (10px) — Dense trees with many nodes
• Medium (20px, default) — General use
• Large (30px) — Presentations and dashboards with few nodes

Tips & Interpretation

1. Order your hierarchy columns carefully: The first column you select becomes the root. A wrong order (e.g., City before Country) produces a nonsensical tree where leaf nodes appear at the root.

2. Left-to-Right orientation for long labels: Role names, file paths, and taxonomy terms are often long. Switching to Left to Right gives each label more horizontal room and prevents overlap.

3. Use Value/Size to add a second story: A tree showing department structure becomes much richer when node size reflects headcount — you see both who reports to whom and how big each team is.

4. Color By categorical = instant grouping: Color nodes by Department Type or Status and the tree gains a second dimension of meaning without changing its layout.

5. Reingold-Tilford keeps siblings evenly spaced: This means the horizontal spread of the tree reflects the number of children at each level, not the values. Wide branches = many children, not large values.

6. Tree Plot vs Treemap: Treemaps encode proportion through rectangle area. Tree plots encode relationships through edges. Use Tree Plot when the connections matter; use Treemap when the sizes matter.

7. Tree Plot vs Sunburst: Sunburst uses radial nesting to show hierarchy. Tree Plot uses explicit edges and node positions. Tree Plot is clearer for deep, narrow hierarchies; Sunburst is better for wide, shallow ones.

8. Tree Plot vs Icicle Chart: Icicle charts also show hierarchy with connected rectangles but without explicit edges. Tree Plot is better for org-chart-style communication; Icicle is better for proportional drill-down.

9. Filter before plotting large trees: Trees with hundreds of leaf nodes become unreadable at the default zoom level. Filter to a single branch or top-N children per level for clarity.

10. Decision tree output: When visualizing a trained decision tree model, map each split column as a hierarchy level and use the node's predicted class or impurity as the Color By column to immediately highlight which branches are most impure or dominant.