AI-Driven Data Visualization Tool, Evaluation

Overview

A local install of Microsoft’s open-source Data Formulator, a tool that uses LLMs to transform data and author rich visualizations from natural-language intent and UI interaction. Set up in a Python virtual environment for evaluation.

Why It Exists

To evaluate concept-driven, AI-assisted visualization: describing the chart you want (and the data shaping it needs) in natural language and letting the LLM generate the transformation and visualization, rather than hand-writing queries and chart specs.

What We Built

No custom code, the upstream package installed locally (Python venv) and run hands-on. Data Formulator pairs a Python/Flask backend with a React UI and an LLM to translate intent into data transformations and charts.

Technologies & Approach

Python/Flask + React with LLM integration for the transform-and-visualize loop. Reviewed for its blended UI + natural-language interaction model.

Outcome / Impact

Gave the studio practical exposure to AI-assisted, natural-language data visualization, complementing related auto-analytics evaluations. Documented as evaluation/R&D.

Capabilities Demonstrated

• Evaluating LLM-driven data-visualization tooling
• Studying natural-language-to-chart interaction patterns