prodigyflow

🚀 ProdigyFlow — Intelligent Data Analytics Agent

ProdigyFlow is an AI-powered multi-agent data analytics framework designed to automate end-to-end exploratory analysis, visualization, insights generation, and reporting. It enables students, analysts, and researchers to turn raw datasets into meaningful business insights with minimal manual effort.

This project has been developed as part of an academic capstone initiative by: Komal Harshita & Priyamvadha Sahasvi Nune,
Department of Computer Science & Engineering.

---

Project Motivation — Why We Chose This

In real-world business environments, analysts spend 70–80% of time cleaning, exploring, and summarizing data before any modelling or decision-making. This process is repetitive, time-consuming, and prone to error.

We wanted to build a system that:

• Speeds up exploratory analysis
• Automatically generates insights and visual summaries
• Allows anyone to analyze datasets without advance ML knowledge
• Demonstrates practical learning combining Python + Data Analytics + Gemini AI

ProdigyFlow reflects our goal to create simple, useful, modular tools that solve real analytical problems while being easy for students and businesses to adopt.

---

🎯 Objectives

• Build a backend-first analytics automation system using agent architecture.
• Support data ingestion, cleaning, visualization, metadata extraction, and insight generation.
• Utilize Gemini models for natural-language summaries and understanding.
• Create a structured, competition-ready engineering pipeline.
• Deliver high-value analytical output with minimal coding.

---

🧠 Core Features

Feature	Description
Automated Data Analysis Agent	Generates insights, metadata & summaries
Visualization Agent	Creates automated charts & visual summaries
Gemini-powered AI Summary	Natural-language insights from data
Structured Output Formatting	Clean and professional console reporting
Modular Agent Design	Add or replace agents independently
CSV/Excel Ingestion Support	Easily test custom datasets

---

🤖 Core System Agents

Agent Name	Responsibility	Output
analysis_agent.py	Reads dataset, extracts statistics, generates Gemini summary	Insights, metadata JSON
visualization_agent.py	Generates visual graphs and saves locally	PNG charts
cleaning_agent.py	Cleans missing values, formatting, and structure	Cleaned dataset
test_gemini.py	Tests Gemini API connection	Model response output

---

🧪 Running the Project Locally

1️⃣ Create and activate virtual environment

python -m venv .venv
.\.venv\Scripts\activate         
source .venv/bin/activate   

2️⃣ Install dependencies

pip install -r requirements.txt

3️⃣ Run the Analysis Agent

python agents/analysis_agent.py

4️⃣ Test Gemini API

python agents/test_gemini.py

---

Add a Custom CSV file

Place your dataset inside the data/ folder and update path in code:

file_path = "data/your_file.csv"

---

Sample Terminal Output Preview

🚀 Running a dry test of analysis_agent...

📂 Using file: data/student_marks.csv

📊 INSIGHTS (Structured Data Overview)
------------------------------------
{ ... dataset overview JSON ... }

🤖 AI-GENERATED SUMMARY
-----------------------
• Key performance trends detected  
• Distribution shows variation in subject performance  
• Potential improvement insights  

📁 METADATA
-----------
{ ... summary JSON ... }

✔ Analysis completed successfully!

---

✨ What We Learned

• Designing modular agent architectures
• Integrating Gemini AI with Python backends
• Automating analytics workflows like real data analysts
• Practical environment setup, dependency management & debugging
• Dataset quality, structure, visualization and reporting best practices
• Leveraging GitHub workflow and documentation standards

---

Future Scope

🔹 Build a web-based interface using FastAPI/Streamlit 🔹 Add database integration and Auto-EDA dashboards 🔹 Support PDF report generation 🔹 Multi-file dataset comparison 🔹 Plug-and-play Machine Learning agent

ProdigyFlow is only the beginning — we plan to expand it into a fully intelligent analytical automation assistant.

---

🤝 Contributors

Name	Role
Komal Harshita	Lead Developer, Agent Architecture, AI Integration
Priyamvadha Sahasvi Nune	Data Research, Analytics, Testing & Documentation

---

🌟 Support

If you like this project, please ⭐ star the repository and share feedback!

🔗 Useful Links

📦 Repository https://github.com/komalharshita/prodigyflow

📘 Project Documentation https://komalharshita.github.io/prodigyflow/

---

📝 License

This project is released under MIT License — feel free to use or modify with attribution.

---

Thank you for exploring ProdigyFlow!

This site is open source. Improve this page.