> ## Documentation Index
> Fetch the complete documentation index at: https://docs.purna.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Visualizations

> Generate charts and scientific figures from data directly in chat.

Purna can generate charts and figures inside a conversation. Ask for a visualization from uploaded data, a referenced file, a table in the conversation, or database-derived results.

## How it works

Attach or reference the relevant data, then describe the figure you want. Purna can create simple inline charts or use [Code Execution](/code-execution) for custom plots, statistics, and publication-style figures.

Examples:

```
"Create a volcano plot from @deseq2_results.csv."
"Make a heatmap of the top 50 variable genes in @expression_matrix.tsv."
"Show a bar chart comparing pathway enrichment scores."
"Plot the distribution of allele frequencies in this table."
"Create a multi-panel figure from these QC metrics."
```

## Supported chart types

| Chart type               | Best for                                              |
| ------------------------ | ----------------------------------------------------- |
| **Bar / horizontal bar** | Comparing categories and ranked values                |
| **Line / area**          | Trends, trajectories, and ordered data                |
| **Scatter**              | Relationships between two variables                   |
| **Pie / donut**          | Simple proportions                                    |
| **Heatmap**              | Matrix-style expression or feature intensity data     |
| **Volcano plot**         | Differential expression or effect-size analysis       |
| **Custom figures**       | Multi-panel scientific figures through code execution |

## Adjusting charts

After Purna generates a chart, ask it to modify the output in the same conversation:

* Change the chart type: *"Make this a horizontal bar chart instead"*
* Update colors: *"Use a colorblind-safe palette"*
* Add or remove data: *"Include only adjusted p-value \< 0.05"*
* Change labels or titles: *"Rename the y-axis to log2 fold change"*
* Export the result: *"Save this as an SVG and PNG"*

<Tip>
  Use simple inline charts for quick summaries. Use [Code Execution](/code-execution) when you need heatmaps, volcano plots, multi-panel figures, statistical annotations, or precise figure formatting.
</Tip>