student performance insight system

Turn raw student score sheets into smart insights with ADK — an AI-powered web assistant that ranks, analyzes, and visualizes academic performance in seconds.

Comment

Inspiration

As a student, I’ve often noticed how academic performance data like marksheets or Excel files just sit unused. Teachers are busy, and students rarely get personal feedback unless it’s exam season. I wanted to create something simple but smart — a system that can take raw data and turn it into useful insights in seconds.

I found out about this hackathon only two days before the deadline, but I decided to participate anyway. I treated it as a way to learn something new and build a complete project from scratch using ADK.

What it does

The ADK Student Insight System is a web-based tool that helps analyze student performance using natural language queries. You just upload a CSV or Excel file with scores, and the system returns:

  • The top student and a full ranking
  • Subject-wise average scores
  • Feedback for each student
  • A bar chart showing subject performance
  • All presented in a clean, readable Markdown report

How I built it

  • I used Google’s Agent Development Kit (ADK) to build a multi-agent system — each agent handles a specific task like reporting, charting, or feedback.
  • The backend is written in Python, using pandas for data analysis and matplotlib for chart generation.
  • The project runs entirely on ADK’s web interface, making it interactive and accessible without setup.
  • I added file upload support, bar chart generation, and clean Markdown formatting for better presentation.

Challenges I ran into

  • Understanding how to structure tools and agents properly in ADK took some trial and error.
  • Working with base64-encoded files and ensuring compatibility with both CSV and Excel was tricky.
  • Making the report clear, visual, and easy to read while keeping the logic modular took careful planning.

Accomplishments I’m proud of

  • Created a working multi-agent system using ADK’s web platform
  • Cleanly separated tasks across agents and utilities
  • Delivered a smooth user experience: upload a file, ask a question, get insights instantly
  • Integrated data analysis, visualization, and feedback generation in one pipeline

What I learned

  • How to build, test, and deploy ADK agents effectively
  • Real-world usage of Python data analysis libraries
  • The importance of clarity in user-facing outputs
  • That even small tools can make raw data meaningful

What’s next

  • Add grading scale support (A/B/C, Pass/Fail, etc.)
  • Export reports as downloadable PDFs
  • Enable question filtering or subject-specific analysis
  • Continue refining UI prompts and multi-agent coordination

Built With

Share this project:

Updates