test changed

Dataset Detective

---

Dataset Detective is a production-grade web application for automated, visual Exploratory Data Analysis (EDA) on any CSV file
Designed for rapid insight, it features a glassmorphic interface, micro-interactions, and exportable analysis reports.
Access the live application: dataset-detective.vercel.app

---

Approach

As an AI/ML learner, my focus is on automating the EDA workflow.
Dataset Detective enables instant, code-free analysis of any dataset, surfacing key metrics, correlations, and distributions.
The platform is engineered for reliability, clarity, and performance, supporting both quick exploration and in-depth reporting.

---

Features

• Universal CSV parsing with tolerant type inference and column unioning
• Missingness metrics per column and overall completeness
• Numeric statistics: count, mean, median, standard deviation
• Categorical insights: unique values and frequency analysis
• Correlation matrix with top relationships
• Per-numeric histograms for distribution visualization
• Glassmorphic UI with animated transitions and accessibility support
• One-click export of interactive analysis as PNG (cross-origin safe)
• Optimized rendering and memoized computations for stability

---

Usage

Visit dataset-detective.vercel.app
Upload your CSV file and immediately explore your data with automated EDA tools.
Export your analysis as a PNG report for sharing or documentation.

---

Project Structure

app/           App Router pages and layout
components/    Feature components (upload, EDA phases, charts, etc.)
components/ui/ shadcn/ui primitives
public/        Static assets (icons, images)
scripts/       Utility scripts (detective.py for offline EDA)
styles/        Global CSS

---

Implementation Notes

• Export uses html-to-image with skipFonts: true and a safe font stack to avoid cross-origin font errors.
• Numeric detection accepts numeric strings; correlations computed only from valid rows; missingness computed against total rows.
• Accessibility: semantic landmarks, aria labels, keyboard focus support, and tabular numerals for key metrics.

---

Customization

• Branding: Update tokens and typography in app/globals.css and titles/icons in app/layout.tsx.
• Icons: Replace /public/icon.jpg and /public/apple-touch-icon.jpg with your logo.
• Splash & Greeting: Tweak timings and copy in components/splash.tsx and components/welcome-greeting.tsx.

---

Deployment

• Vercel: Deploy directly from GitHub. No special environment variables required for base CSV workflow.

---

Troubleshooting

• PNG export fails with “cssRules” error:
  • Ensure skipFonts: true remains set in the export function.
  • Retry after closing browser extensions that inject styles.
• Styling issues:
  • Confirm Tailwind v4 is active and your tokens are present in globals.

---

License

MIT

---

Author

Sai
AI/ML Learner
GitHub: sai-AIstacker
Live App: dataset-detective.vercel.app

---