[🇧🇷 Português] [🇬🇧 English]

16- Data Mining / LLM Tabular Preprocessing with Dictionary Groups

Dictionary-Based Feature Grouping for LLM/AI Pipelines

Institution: Pontifical Catholic University of São Paulo (PUC-SP)
School: Faculty of Interdisciplinary Studies
Program: Humanistic AI and Data Science Semester: 2nd Semester 2025
Professor: Professor Doctor in Mathematics Daniel Rodrigues da Silva

Note

⚠️ Heads Up

• Projects and deliverables may be made publicly available whenever possible.
• The course emphasizes practical, hands-on experience with real datasets to simulate professional consulting scenarios in the fields of Data Analysis and Data Mining for partner organizations and institutions affiliated with the university.
• All activities comply with the academic and ethical guidelines of PUC-SP.
• Any content not authorized for public disclosure will remain confidential and securely stored in private repositories.

🎶 Prelude Suite no.1 (J. S. Bach) - Sound Design Remix

Statistical.Measures.and.Banking.Sector.Analysis.at.Bovespa.mp4

📺 For better resolution, watch the video on YouTube.

Tip

This repository is a review of the Statistics course from the undergraduate program Humanities, AI and Data Science at PUC-SP.

☞ Access Data Mining Main Repository

📚 Table of Contents

1. Overview
2. What is Dictionary-Based Feature Grouping?
3. Why Use This Technique?
4. Key Concepts
5. Installation
6. Quick Start
7. Basic Examples
8. Advanced Usage with LLMs
9. Real-World Applications
10. Project Structure
11. Notebooks
12. Dataset Resources
13. References
14. Contact

Overview

This repository demonstrates dictionary-based feature grouping for tabular data preprocessing, specifically designed for integration with Large Language Models (LLMs) and AI/ML pipelines.

The technique allows you to organize related columns (features) in a dataset using dictionaries, enabling:

• Semantic grouping of features
• Efficient preprocessing for LLM-based feature engineering
• Better interpretability of tabular data
• Streamlined data transformation pipelines

Tip

Perfect for: Data Scientists, ML Engineers, AI Researchers, and Students working with tabular data and LLMs!

What is Dictionary-Based Feature Grouping?

Dictionary-based feature grouping is a data preprocessing technique where you use Python dictionaries to organize and group related columns (features) in a DataFrame based on their semantic meaning or data type.

💡 Simple Explanation (For Beginners)

Imagine you have a dataset about customers with many columns:

age, income, city, state, country, purchase_date, product_name, price, ...

Instead of processing all columns individually, you can group them by meaning:

feature_groups = {
    'demographics': ['age', 'income'],
    'location': ['city', 'state', 'country'],
    'transaction': ['purchase_date', 'product_name', 'price']
}

This makes it easier to:

1. Apply specific transformations to each group
2. Feed organized data to LLMs
3. Understand your dataset structure
4. Create modular and maintainable code

Why Use This Technique?

For Traditional ML

• 📦 Organized Feature Engineering: Group numerical, categorical, and text features separately
• ⚛️ Pipeline Efficiency: Apply different transformers to different feature groups
• 🧠 Better Understanding: Know which features belong together conceptually

For LLM Integration

• 🤖 Semantic Context: LLMs perform better when features are semantically grouped
• 💬 Prompt Engineering: Create structured prompts with organized feature groups
• 🔗 Hybrid Models: Combine tabular data with LLM embeddings effectively
• 🚀 Feature Generation: Use LLMs to create new features from grouped columns

📝 Key Concepts

1. Pandas GroupBy

Core Python/Pandas functionality for splitting, applying, and combining data:

df.groupby('category').agg({'value': 'mean'})

2. Dictionary Mapping

Using dictionaries to define feature relationships:

column_mapping = {
    'group_name': ['col1', 'col2', 'col3']
}

3. LLM Feature Engineering

Leveraging LLMs to:

• Generate text embeddings from grouped text columns
• Create semantic features
• Enrich tabular data with contextual information

Installation

Prerequisites

• Python 3.8+
• pip or conda

Install Dependencies

# Clone the repository
git clone https://github.com/Quantum-Software-Development/16-DataMining_llm-tabular-preprocessing-dict-groups.git
cd 16-DataMining_llm-tabular-preprocessing-dict-groups

# Install requirements
pip install -r requirements.txt

Docker Setup (Optional)

# Build Docker image
docker build -t dict-groups-preprocessing .

# Run container
docker run -p 8888:8888 dict-groups-preprocessing

Quick Start

Basic Example

import pandas as pd

# Sample dataset
data = {
    'name': ['Alice', 'Bob', 'Charlie'],
    'age': [25, 30, 35],
    'city': ['NY', 'LA', 'Chicago'],
    'salary': [50000, 60000, 70000],
    'department': ['IT', 'HR', 'IT']
}
df = pd.DataFrame(data)

# Define feature groups
feature_dict = {
    'personal': ['name', 'age'],
    'location': ['city'],
    'professional': ['salary', 'department']
}

# Process by group
for group_name, columns in feature_dict.items():
    print(f"\nProcessing {group_name}:")
    print(df[columns].head())

Output:

Processing personal:
      name  age
0    Alice   25
1      Bob   30
2  Charlie   35

Processing location:
       city
0        NY
1        LA
2   Chicago

Processing professional:
   salary department
0   50000         IT
1   60000         HR
2   70000         IT

Basic Examples

Example 1: Grouping by Data Type

import pandas as pd
import numpy as np

# Create sample data
df = pd.DataFrame({
    'id': [1, 2, 3],
    'name': ['Product A', 'Product B', 'Product C'],
    'price': [10.5, 20.0, 15.75],
    'category': ['Electronics', 'Clothing', 'Electronics']
})

# Group columns by data type
type_groups = {
    'numeric': df.select_dtypes(include=[np.number]).columns.tolist(),
    'categorical': ['category'],
    'text': ['name']
}

🤖 Advanced Usage with LLMs

LLM-Based Feature Generation

# Example: Using grouped text features for LLM prompts
text_groups = {
    'product_info': ['product_name', 'description'],
    'user_feedback': ['reviews']
}

# Create structured prompt for LLM
def create_llm_prompt(row, group_dict):
    prompt = ""
    for group_name, cols in group_dict.items():
        prompt += f"{group_name}: {', '.join([str(row[col]) for col in cols])}\n"
    return prompt

br>

🌐 Real-World Applications

1. E-commerce: Group product features, pricing, and reviews
2. Healthcare: Organize patient demographics, vitals, and medical history
3. Finance: Separate transaction data, customer info, and risk factors
4. NLP: Combine tabular + text data for hybrid models

📂 Project Structure

16-DataMining_llm-tabular-preprocessing-dict-groups/
│
├── Codes/
│   ├── notebooks_01_basic_example.ipynb
│   └── notebooks_02_llm_preprocessing.ipynb
│
├── requirements.txt
├── requirements-dev.txt
├── requirements-full.txt
├── requirements-lock.txt
├── pyproject.toml
├── .gitignore
├── LICENSE
├── README.md
└── README.pt_BR.md

br>

Notebooks

1. notebooks_01_basic_example.ipynb

• Introduction to dictionary-based grouping
• Basic Pandas operations
• Simple examples with sample data

2. notebooks_02_llm_preprocessing.ipynb

• Advanced LLM integration
• Feature generation using grouped data
• Real-world dataset examples

👉 Open in Colab: Basic Example | LLM Preprocessing

Dataset Resources

The notebooks use publicly available datasets:

• UCI Machine Learning Repository: https://archive.ics.uci.edu/ml/index.php
• Kaggle Datasets: https://www.kaggle.com/datasets
• Hugging Face Datasets: https://huggingface.co/datasets

References

1. Chen, X., et al. (2024). LLM-based feature generation from text for interpretable machine learning. arXiv preprint. Retrieved from arxiv.org/html/2409.07132v2

2. DataCamp. (2024). Pandas GroupBy Explained: Syntax, Examples, and Tips. Retrieved from datacamp.com/tutorial/pandas-groupby

• GeeksforGeeks. (2024). Pandas dataframe.groupby() Method. Retrieved from geeksforgeeks.org

• Machine Learning Mastery. (2024). Feature Engineering with LLM Embeddings: Enhancing Scikit-learn Models. Retrieved from machinelearningmastery.com

• McKinney, W. (2017). Python for Data Analysis: Data Wrangling with Pandas, NumPy, and IPython (2nd ed.). O'Reilly Media.

• Pandas Documentation. (2024). Group by: split-apply-combine. Retrieved from pandas.pydata.org/docs/user_guide/groupby.html

• VanderPlas, J. (2016). Python Data Science Handbook: Essential Tools for Working with Data. O'Reilly Media.

💌 Let the data flow... Ping Me !

🛸๋ My Contacts Hub

────────────── 🔭⋆ ──────────────

➣➢➤ Back to Top

Copyright 2026 Quantum Software Development. Code released under the MIT License license.