Complete Chart Gallery

Choose the question before the chart

Pick a chart that answers your decision question clearly. Every card here shows the chart, when to use it, when to avoid it, and the design references it draws from.

The future of business analytics is AI-infused: analysts express intent, AI helps shape evidence and visuals, and people can still inspect, challenge, and defend the conclusion.

Choose by question
  • Compare categories? Bar family — Bar, Grouped, Stacked, Horizontal, Lollipop.
  • Change over time? Line for trend; Area when magnitude matters; Combo for two scales.
  • Show a distribution? Histogram for shape; Box/Violin/Ridgeline for group-vs-group.
  • Show a relationship? Scatter, Bubble, Regression, Connected Scatter.
  • Show composition? Pie/Donut only for 2-3 slices; ranked bar for more; Sunburst for hierarchies.
  • Show patterns in dense data? Heatmap, Sankey, Streamgraph.
  • Communicate a single value? KPI Card, Bullet, Sparkline.
Design references we lean on

Cole Nussbaumer Knaflic — Storytelling with Data. Every chart carries a one-sentence Big Idea; eliminate clutter; direct attention with pre-attentive attributes.

Andy Kirk — Data Visualisation. Trustworthy, accessible, elegant; perceive, interpret, comprehend.

Stephen Few — Show Me the Numbers & Information Dashboard Design. Tables interact with the verbal system; charts with the visual. Bullet beats Gauge. Fit one screen.

Wexler, Shaffer & Cotgreave — Big Book of Dashboards. Small multiples turn a spaghetti chart into a scan; test with real users.

Gestalt principles. Proximity, similarity, enclosure, connection, continuity — how the eye reads groupings.

Accessibility. Every professional chart in this gallery ships from a colourblind-safe palette. Charts marked with an accessibility indicator also need direct labels, patterns, or high-contrast alternatives.

How the badges work

PBI — Native Power BI visual, available in Power BI Desktop out of the box.

AppSource — Available as a marketplace visual for Power BI or equivalent tooling elsewhere.

Family headers and the sticky navigation establish the analytical family. The card subtitle states the chart's Big Idea; the badge beside its name identifies its implementation status.

About Flint Charts

Flint is a visualization language built by Microsoft Research in collaboration with the IDEAS Lab, Renmin University of China. It turns compact semantic chart specifications into polished, backend-native visualizations.

This Gallery uses flint-chart to compile the same semantic input into Vega-Lite charts. The companion flint-chart-mcp server makes that pipeline available to AI agents for local chart creation, validation, and rendering.

Data Formulator, also a Microsoft Research project, shows how AI-agent data exploration can build on Flint: analysts can work from their data, branch into alternative questions, and refine visual evidence without losing the analytical thread.

This is the future of AI-infused business analytics: AI accelerates the work, while human judgment stays accountable for the decision.

Gallery navigation and presentation styles