DataLens

CSV analytics dashboard. Upload any CSV file and instantly get auto-generated charts, summary statistics, and AI narrative insights — no coding required.

Built for CS 433 — Data Analytics at Southeast Missouri State University.

Features

• Drag and drop CSV upload (or browse to select)
• KPI cards: row count, column count, numeric columns, missing values
• Auto-generated bar chart (first numeric column)
• Auto-generated doughnut chart (first categorical column)
• Data preview table (first 8 rows, scrollable)
• AI narrative insights: summary, 4 key findings, anomaly detection, recommendation
• 3 built-in sample datasets (sales, student grades, tech jobs)
• Reset and upload a new file without refreshing

Project Structure

datalens/
├── index.html        # HTML layout and structure
├── css/
│   └── style.css     # Dark purple theme, CSS variables, responsive grid
├── js/
│   └── app.js        # CSV parsing, chart building, AI insights, state management
└── README.md

Tech Stack

• Vanilla HTML / CSS / JavaScript
• Chart.js 4.4.1 — bar and doughnut charts
• PapaParse 5.4.1 — CSV parsing
• LLM inference API — natural language dataset insights
• Fonts: Outfit + Fira Code (Google Fonts)

Running Locally

Open index.html in any modern browser. Chart.js and PapaParse load from CDN — internet connection needed.

AI insights require a valid API key configured in js/app.js.

Known Issues / TODO

• Wide CSVs (20+ columns) cause horizontal scroll in the preview table on mobile — need to add column pinning
• Doesn't support Excel files (.xlsx), CSV only for now
• TODO: add chart export to PNG

Course Context

In CS 433 we spent a lot of time on exploratory data analysis in Python — loading datasets, checking dtypes, plotting distributions. This project automates that first-pass EDA step for non-technical users. The AI layer surfaces patterns that would otherwise take several minutes of manual analysis to spot.

Author

Tarunima Amisha · github.com/amisha53 · SEMO CS 2026