Real-World Data Cleaning

Author: Uttarayan Haldar Objective: Practicing real-world inspired data cleaning challenges daily to master preprocessing, wrangling, and transformation for Data Science & Analytics.

Overview

This repository contains my daily data cleaning challenges inspired by real-world datasets. Each challenge simulates messy, inconsistent, or incomplete data, and I document the process of transforming it into analysis-ready form.

---

Skills & Tools

• Languages: Python (Pandas, NumPy)

• Libraries: Faker, datetime, re, pyjanitor

• Techniques:

  • Handling missing values
  • Removing duplicates
  • Fixing data types
  • String/text cleaning
  • Outlier detection & treatment
  • Feature transformation & encoding

Daily Challenge Structure

Each day’s folder contains:

1. dataset/ – Raw and cleaned CSV files.
2. notebook/ – Jupyter Notebook documenting the cleaning process.
3. notes.md – Summary of learnings from the day’s challenge.

Learning Goals

• Build muscle memory for fast, accurate cleaning.
• Simulate real-world messiness (typos, mixed formats, outliers).
• Develop reproducible cleaning scripts for portfolio use.

Connect

Email: uttarayan.haldar.data@gmail.com