WIP: Agentic Workflow for Automated Data Visualization using LangChain

[sayantikabanik]

About this PR

This PR introduces an automated data visualization framework leveraging LangChain, GitHub AI models, and agentic design principles to suggest and generate visualizations with minimal human intervention.

What is the Agentic way of designing in short

Focus on autonomy, adaptability, and proactive decision-making.
Instead of traditional rigid workflows. Agentic workflows are equipped to make informed choices based on changing conditions.

Key Features:

• Visualization Suggestion: Determines the appropriate visualization type based on the data characteristics.
• Draft Generation: Creates draft visualizations using Matplotlib for quick insights.
• LangChain Integration: Utilizes GitHub AI models via LangChain for intelligent decision-making in visualization tasks.

Workflow Details:

• Autonomous Suggestions: Based on the data provided, the framework suggests suitable
• To be worked upon: visualization types (e.g., histograms, scatter plots, bar charts).
• @langchain-ai Agent: An agent initialized with tools for suggesting visualizations and generating drafts, enhancing adaptability and proactive decision-making.

[sayantikabanik]

awaiting token renewal for one complete round of testing

[sayantikabanik]

TODO

• modify the request body
• adjust token limit/ check the usage real time
• fix the error log (reduce verbosity)