Plotly plugin

[richardfogaca]

Plotly Plugin for Superset

Contribution Summary

New plugin that brings Plotly dynamic visualizations to Superset.

What We Added

We created a new visualization plugin that lets you use Python and Plotly to create custom charts in Superset. This plugin:

1. Adds a Python code editor directly in Superset's UI
2. Executes Python code on the server to generate Plotly visualizations
3. Renders interactive Plotly charts in Superset dashboards

How It Works

Frontend

• A code editor with Python syntax highlighting lets you write Plotly code
• The standard Superset query builder provides data to your visualization
• The Plotly chart renders using react-plotly.js for optimal performance

Future Improvements (Planned)

• Enhanced error handling with user-friendly error messages
• Syntax validation before submission to prevent server-side errors
• Code templates and snippets for common visualization types

Backend

• Your Python code runs on the server with access to the query results
• Security measures prevent dangerous code execution
• The generated Plotly JSON is sent to the frontend for rendering

Security Enhancements (Planned):

• Execution time limits
• Restrict certain imports for security reasons
• Expanded allowlist for safe imports (pandas, numpy, etc.)

Screenshot

Example Code

Here's a simple example of what you can write in the code editor:

import plotly.graph_objects as go
import plotly.express as px

def get_plotly(df):
    # Create a bar chart using the query results
    fig = px.bar(
        df,
        x="category",
        y="revenue",
        color="region",
        title="Revenue by Category and Region"
    )
    
    # Customize the layout
    fig.update_layout(
        legend_title="Region",
        xaxis_title="Category",
        yaxis_title="Revenue ($)"
    )
    
    return fig

Why This Matters

This plugin solves several problems:

1. Visualization Limitations: Standard Superset charts don't cover all use cases
2. Custom Analytics: Some analyses require specialized visualizations
3. Python Workflow: Data scientists often work in Python but need to share results
4. Dynamic Visualizations: Unlike standard ECharts-based visualizations, Plotly charts can be completely customized with Python code, enabling dynamic graphics that adapt to the data

Limitations

• Cross-Filter: The Plotly plugin currently doesn't support Superset's cross-filtering capabilities (where clicking on one chart filters other charts in the dashboard).

[lilila]

This is a Wonderfull functionality, thanks for sharing

[rusackas]

Is it available somewhere to install or preview? Looks cool!

[richardfogaca]

Hi @rusackas , I'm glad you liked the idea. I implemented a test version on this [fork] (https://github.com/richardfogaca/superset/tree/plotly-prototype)

I'm also attaching a chart I created. If this feature is interesting for superset please just let me know :)

chart_export_20250401T130236.zip

[rusackas]

I love the idea of adding more... the only issue would be that of supporting/maintaining them, where we're already still building/improving/fixing ECharts and haven't completed all the migrations from legacy charts yet.