Text2Insight — Natural Language Analytics Platform

GitHub ↗

The Business Problem

Business teams depend on analysts or engineers for even simple data questions, creating delays and bottlenecks.

The Solution

A natural-language analytics platform — ask a question in plain English, get a live chart back in seconds.

The Result

Query Response Time

→ 3-8 sec

SQL Generation Accuracy

→ ~90%+ on standard queries

Manual Analysis Time

15-30 min → Seconds

Self-reported from my own testing, not a third-party audit.

Removes SQL dependency for everyday business questions
Automated chart generation alongside every answer
One-click PDF / Word / PPT export for sharing results
Multi-turn conversation memory supports follow-up analysis
Admin observability dashboard for monitoring query quality

Architecture

User Question -> LLM SQL Generation (confidence-scored) -> DuckDB Execution -> Chart + Summary (streamed via SSE)

⚙️ Technical Details For engineers & technical reviewers — click to expand

Turns a plain English question into a live chart in seconds — no SQL, no analyst, no waiting. Converts natural language into SQL, executes it against a live sales database, and streams back data tables, charts, and LLM-generated summaries in real time. Built to production-grade standards: JWT authentication, multi-turn conversation memory, confidence-scored SQL generation, PDF/Word/PPT export, and an admin observability dashboard.

FastAPI DuckDB SSE Streaming HTML5 CSS3 Vanilla JS Redis-ready caching

Need something like this for your own business?

Build Something Similar →

Have questions about this project? Ask my AI assistant for details.

Ask AI about this →