Auto-EDA Dashboard — Instant Data Profiling
A live, gated demo: upload a CSV (or a built-in sample) and get an instant exploratory data analysis — column profiling, auto-generated charts, and an AI-written summary of what stands out. LLM-agnostic; this instance uses Gemini.
Overview
Auto-EDA Dashboard turns a raw CSV into an instant analytical dashboard. On upload, every column is profiled (type, null rate, distinct count, min/max/mean/median/standard deviation, statistical outliers, and detected formats), the most informative charts are generated automatically, and an AI analyst writes a short executive narrative of what stands out — all grounded in the computed statistics. The numbers are calculated locally over an in-memory SQLite engine; the LLM only writes the narrative. LLM-agnostic — runs on a local model (Ollama / Llama) or any cloud LLM; this deployment uses Gemini. Data is processed in-memory and expires automatically.
Problem Solved
Understanding a new dataset normally takes an analyst hours of manual profiling and charting. This demo compresses that first pass into seconds — surfacing nulls, ranges, outliers, and distributions with charts and a plain-English summary, so teams know what they are working with before they invest in a pipeline.
Capabilities
- CSV upload or built-in sample datasets
- Per-column profiling: types, nulls, distinct, min/max/mean/median/stddev
- Statistical outlier detection (IQR)
- Text format detection (email, date, phone, URL, …)
- Auto-generated bar/line charts
- AI-written executive summary grounded in the stats
- In-memory processing, rate limiting, gated access