Smart Data Clean & Query Agent

An AI-powered Streamlit app that automates dataset cleaning, data type casting and allows natural language querying using Google Gemini AI.

Comment

Inspiration

Data preprocessing, missing value treatment, and manual schema checking take up significant time during data analysis. I created the Smart Data Clean & Query Agent to provide an intuitive, all-in-one solution that automatically cleans raw datasets, handles missing values, and enables users to interact with their data using natural language queries.

What it does

  • Automatic Data Cleaning: Detects and corrects column data types, handles null values intelligently (mean for numeric, 'Unknown' for text), and removes duplicate entries.
  • Schema & Health Diagnostics: Displays missing value counts, dataset shape, and missing percentage summaries instantly.
  • AI-Powered Data Querying: Integrates Google Gemini AI (gemini-1.5-flash) to answer analytical and natural language questions directly from the dataset.
  • Export Ready: Allows users to download the processed and cleaned dataset as a CSV file in a single click.

How we built it

  • Frontend & Framework: Streamlit (Python)
  • Data Processing: Pandas & NumPy
  • AI Model Integration: Google Generative AI SDK (google-generativeai)
  • Deployment: GitHub & Streamlit Community Cloud

Challenges we ran into

  • Designing an efficient pipeline for dynamic type conversion and dataset summary generation.
  • Ensuring secure API Key management by allowing users to enter their Google Gemini API key safely via the sidebar interface.

Accomplishments that we're proud of

  • Successfully deployed a fully functional, user-friendly data assistant using Streamlit and Gemini API.
  • Implemented real-time dynamic query answering directly based on uploaded dataset structure and summary.

What we learned

  • Deepened hands-on knowledge of Streamlit app deployment and custom UI styling.
  • Learned best practices for prompt structuring when feeding dataset metadata into Large Language Models (LLMs).

What's next for Smart Data Clean & Query Agent

  • Adding automated data visualization charts (using Plotly/Seaborn).
  • Supporting multi-file upload and SQL database connection capabilities.

Built With

Share this project:

Updates

Submission history