DataSense-AI

📊 DataSense AI - Dataset & SQL Copilot

Live demo: https://datasense-si.streamlit.app/ · Product portfolio: https://linda8-lang.github.io/DataSense-AI/portfolio.html

DataSense AI turns raw spreadsheets into executive-ready insights. Upload a dataset, ask a question in plain English, and get a short summary plus a ready-to-run SQL query.

The Problem

Data analysts lose hours translating raw files into executive summaries and SQL queries. General-purpose LLMs hallucinate numbers and can’t do exact computation, and non-technical teams struggle to write good prompts.

How DataSense Solves It

• Exact math with Pandas - row counts, column counts and missing values are computed directly from the data, never guessed by the AI.
• AI for language, not arithmetic - the LLM only sees the column names and types, and writes the plain-English summary and the ANSI SQL query.

Features

• 📁 Flexible data sources - CSV, Excel (.xlsx, .xls, with a sheet picker) and public Google Sheets links
• 🔍 Instant data profile - preview, total rows, total columns and missing values
• 💬 Text-to-SQL - ask a question in plain English and get a 2-sentence summary plus an optimized SQL query
• ⚡ Fast and free - runs open-source models on Groq’s free tier

Tech Stack

| Layer | Tool | |——-|——| | UI | Streamlit | | Data engine | Pandas, openpyxl, xlrd | | LLM | openai/gpt-oss-20b via Groq | | Config | python-dotenv |

Run It Locally

git clone <this-repo-url>
cd DataSense
pip install -r requirements.txt
cp .env.example .env      # then paste your free Groq key into .env
streamlit run app.py

Get a free Groq API key (no credit card) at https://console.groq.com.

Project Docs

See PRD.md for the product requirements, user persona and KPIs.