Supper Charts: Natural language data visualization
Data visualization is fundamental to understanding data during analysis and communicating what the data means afterward. Before Supper added native charting, users could retrieve and analyze data but had to take the results elsewhere to visualize them. That meant downloading data, copying it into another application, or rebuilding the analysis inside a spreadsheet or traditional business intelligence tool.
Supper Charts brings analysis and visualization into one continuous workflow. Users can create charts from analytical results using natural language, then continue refining those charts conversationally.
What Is Supper Charts?
Supper Charts is Supper's native data visualization capability. It lets users turn data produced during an analytical conversation into a chart, then modify the visualization by telling Supper what they want to change. A user can request a specific chart type or accept a visualization suggested by the agent.
Charts support two distinct types of data work:
- Exploratory analysis: One-time visualizations created while investigating a question in chat
- Persistent reporting: Visualizations used as widgets in dashboards for recurring communication and monitoring
During exploration, a chart can help someone recognize a trend, compare groups, detect an unusual result, or decide which question to ask next. When an insight needs to be monitored or shared repeatedly, the visualization can become part of a persistent dashboard. This connects one-time investigation with ongoing reporting without requiring users to reconstruct the work in another system.
Traditional spreadsheets and BI tools usually expose chart creation through menus, dropdowns, formatting panels, axis controls, and configuration screens. These interfaces can be powerful, but sophisticated changes often require users to understand how a particular chart editor is organized.
Supper uses a different interaction model. SQL generation, data retrieval, manipulation, and analysis already happen through an agentic workflow, so chart construction and editing should work the same way. Traditional BI asks users to operate the chart editor; Supper asks users to describe the chart they want.
How Supper Charts Works
The workflow has three steps:
- Ask: Retrieve or analyze data by asking Supper a business question.
- Visualize: Request a chart or accept a visualization suggested by the agent.
- Refine: Describe what should change until the chart communicates the intended result.
The chart is generated from data produced during the conversation. Supper's agents can also recognize when visualization would be a useful next step after retrieving data, allowing the user to accept a chart suggestion directly. The same conversational interface used to investigate the data becomes the interface for presenting it.
Question → data → chart → critique → revision
Supper Charts remain editable after their initial generation. Users can request a different visualization or modify individual aspects of an existing chart without rebuilding it from scratch.
For example, a user could say:
Show this as a line chart. Put revenue on the left axis and margin on the right. Remove the incomplete current month because it is misleading.
That instruction can replace a sequence of chart-editor operations. The user describes the desired outcome in business language instead of locating the correct settings for the chart type, axes, series, and date range.
Users can switch between chart types, change which variables appear on each axis, modify axis units, assign series to different Y axes, create dual-axis charts, group or stack series, show values as percentages, combine lines and bars, and exclude misleading periods or data points. Conversational editing allows the visualization to evolve alongside the analysis.
What Visualizations Does Supper Charts Support?
Supper Charts supports a broad range of visualization types:
| Analytical purpose | Supported visualizations |
|---|---|
| Time series and continuous data | Line charts and area charts |
| Categorical comparisons | Bar charts and column charts |
| Composition | Pie charts and donut charts |
| Multi-series comparisons | Grouped, stacked, and percentage-stacked bars |
| Relationships | Scatter plots |
| Geographic analysis | Choropleth maps with United States and global coverage |
| Dimensional analysis | Heat maps and matrix-style views |
| Dashboard summaries | KPI cards and progress bars |
| Combined comparisons | Combination charts, including lines and bars together |
This broad visualization surface is intentional. Business users and data teams already expect meaningful flexibility from BI software, so an AI-native alternative must support more than a handful of basic charts. The differentiator is not that these visualizations exist (traditional charting products already offer extensive capabilities), but how users create and modify them.
Why Agentic Charting Is Technically Different
Chart generation presents a different technical problem from SQL generation. When an agent writes SQL, it can inspect the query, execute it, examine the returned data, and use downstream validation signals. A chart is different because the customer ultimately experiences a visual artifact, while the system initially produces a technical chart specification.
That specification can be technically valid and still generate a chart that is confusing or misleading. Correct code does not automatically produce effective data communication. Supper therefore treats chart generation as a specialized capability with chart-specific instructions, tools, rendering logic, and iteration rather than adding every visualization responsibility to the primary data agent.
This challenge illustrates a larger point about AI-native analytics. Producing technically valid output is not enough; an analytical system must also help users shape that output into something useful, understandable, and appropriate for the intended audience.
From Chart Controls to User Intent
Supper Charts reflects a broader product thesis: AI-native business intelligence should not mean placing a chatbot in front of an unchanged BI workflow. The interaction model of the analytical tools themselves should change.
A traditional charting workflow looks like this:
Select data → select a chart → find the correct settings → configure each element → repeat
Supper's model is different:
Describe the desired result → inspect it → describe the change
Natural language turns chart controls into user intent. A business user can iteratively produce a sophisticated visualization without first learning the configuration system of a BI tool, while experienced data professionals can express detailed changes more directly.
Frequently Asked Questions
What is Supper Charts?
Supper Charts is Supper's native data visualization capability. It lets users create charts from analytical results and revise those charts through natural-language instructions.
How do you create and edit a chart in Supper?
First, ask Supper to retrieve or analyze data. Then request a visualization or accept one suggested by the agent. After the chart is created, use follow-up instructions to change supported properties such as its type, axes, units, series, stacking, or displayed range.
What chart types does Supper support?
Supper supports line, area, bar, column, pie, donut, grouped and stacked bar, scatter, choropleth map, heat-map, matrix, KPI, progress-bar, and combination visualizations.
Can Supper Charts be used in dashboards?
Yes. Supper supports exploratory charts created during analysis as well as persistent visualizations used in dashboards.
How is Supper Charts different from a traditional BI chart editor?
Traditional BI tools generally require users to manipulate chart controls and configuration panels. Supper lets users describe the desired visualization and subsequent changes in natural language.