🧹 AI-Powered Data Cleaning Assistant

A no-code assistant to clean your messy CSV datasets using intelligent profiling and AI-generated code suggestions.

---

📌 Introduction

🔍 The Problem

Data professionals often spend 60–80% of their time just cleaning and preparing data before they can analyze or model it. Datasets frequently include:

• Missing, duplicate, or inconsistent values
• Malformed date formats and encoding issues
• Incorrect data types or string-numeric mismatches
• Nulls, outliers, and redundant columns

This repetitive and error-prone work slows down projects and consumes valuable analyst hours.

🤖 The Solution

The AI Data Cleaning Assistant tackles this problem by:

• Profiling the dataset intelligently
• Identifying issues like nulls, duplicates, and type mismatches
• Generating Python code to fix these issues using an LLM
• Presenting actionable summaries and ready-to-use cleaning code in a user-friendly web interface

Users simply upload a CSV file, and the tool does the rest—no coding required.

---

⚙️ How to Run the Project

🔧 Step 1: Clone the Repository

git clone https://github.com/DSM2499/AI_Data_Cleaner.git
cd AI_Data_Cleaner

🐍 Step 2: Set Up a Virtual Environment

python3 -m venv venv
source venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

🚀 Step 3: Launch the Streamlit App

streamlit run app.py

Upload a CSV file through the interface and receive data insights + cleaning code in seconds.

---

🛠️ Tech Stack & Skills

• Frontend/UI: Streamlit
• Backend: Python 3.11+
• Data Profiling: pandas, pyjanitor
• AI Model: Open AI GPT-3.5-Turbo (can be switched out with a more efficient model)
• Modular Python with CLI and UI separation

---

📸 Screenshots

Landing Page

Basic Summary of Dataset and Action Plan

AI Generated Results

📹 YouTube Tutorial

📺 https://youtu.be/uppVlDoWTQc

---

📈 Impact & Value

• ⏱️ Time Saved: Reduces 60–80% of manual cleaning time.
• 📊 Better Insights: Starts with clean data, leading to more reliable analytics.
• 👥 Team Productivity: Standardizes preprocessing workflows across analysts.
• 🚀 Upskilling: Helps beginners learn cleaning techniques via readable code.
• ⚙️ Automation: Makes data pipelines more efficient and less error-prone

---

🧾 Conclusion

This AI-powered tool bridges the gap between raw data and actionable insights—removing the grunt work and letting analysts focus on strategy and decision-making. It serves as both a learning tool and a real-world automation system for streamlining the data cleaning process.

⭐️ Feel free to fork this repo, suggest enhancements, or open issues for bugs and ideas!