Analytics dashboard prototype that surfaces insight cards, anomaly flags and a conversational query interface on top of tabular data.
Live demo · Built with React, Recharts, Tailwind CSS and Vite
Lens explores how a business user could ask questions of their data without writing queries. It derives insight cards from a retail sales dataset, flags anomalies with a drill-down drawer, keeps a decision trail and answers plain-language questions in a chat bar. Insights (src/lib/insights.js) and queries (src/lib/queryEngine.js) run client-side with no backend or LLM calls; it is a UX prototype of the idea. The sample data is a generated Superstore-style dataset, produced by scripts/generate-data.mjs.
git clone https://github.com/ffmc/lens.git
cd lens
npm install
npm run generate-data # regenerates the sample dataset
npm run dev| Path | Purpose |
|---|---|
src/components/ |
Insight cards, anomaly drawer, decision trail, chat bar |
src/lib/ |
Insight generation and query engine |
src/hooks/ |
Data loading |
scripts/generate-data.mjs |
Sample data generator |