← Files Build Web Data VisualizationARCHIVED FILE
references/chart-selection-patterns.md
2.73 KB · Sep 30, 2026 · 22:49 UTC
# Chart Selection Patterns ## What Problem This Solves This reference maps recurring analytical tasks to reliable chart families. ## When to Use It Use this when the user is deciding between bars, dots, lines, heatmaps, small multiples, tables, maps, or more complex structures. ## Key Takeaways - Use bars and dots for precise category comparison. - Use lines and small multiples for time-based comparison. - Use tables with embedded cues such as bars, dots, and sparklines when exact values or compact row-wise trend comparison matter. - Use Gantt charts when the task is project scheduling: planned start/end spans, milestones, dependencies, predecessors, critical path, baselines, WBS, or resource allocation over time. - Use milestone timelines instead of Gantt charts when the data has only key dates and no meaningful task spans. - Use Kanban, tables, calendars, dependency graphs, resource timelines, or uncertainty views when workflow state, lookup, booking, topology, capacity, or risk distributions matter more than planned schedule bars. - Use UML, C4, ERD, BPMN, flowcharts, swimlanes, state machines, sequence diagrams, or dependency diagrams when the evidence is system structure, process, state, interaction, database schema, code shape, or architecture. - Use scatterplots and density views for relationships. - Use maps only when location is part of the reasoning. ## Common Mistakes - Using pies for precise comparison. - Choosing a map when a sorted bar chart would be clearer. - Choosing a network when grouped tables or matrices would explain more. - Choosing a Gantt chart for issue status or task ownership when dates and dependencies are not the main evidence. - Choosing a class diagram for architecture communication when a C4 container/component view or dependency graph would be clearer. - Choosing a generic flowchart when BPMN semantics, swimlane ownership, or state-machine transition validity is the main point. - Treating created/closed timestamps as planned schedule dates without saying the view is an actual lifecycle timeline. - Pulling the key away from the chart so the viewer has to bounce between evidence and legend. ## Adjacent Skills - `../SKILL.md` - `../../gantt-chart-visualization/SKILL.md` - `../../uml-and-software-architecture-visualization/SKILL.md` - `../../geospatial-and-cartographic-visualization/SKILL.md` ## Source Links - [From Data to Viz](https://www.data-to-viz.com/) - [FT Visual Vocabulary](https://github.com/ft-interactive/visual-vocabulary) - [Datawrapper Academy](https://academy.datawrapper.de/) - [OMG UML 2.5.1](https://www.omg.org/spec/UML/2.5.1/About-UML) - [C4 model diagrams](https://c4model.com/diagrams) - [Atlassian Gantt chart overview](https://www.atlassian.com/agile/project-management/gantt-chart)
SHA-256: 52f189052bad2dbef953c301f7496c0e555aa38890c99573e420e508356fca4c