SQL Data Cleaning – World Layoffs Dataset

This project demonstrates how I used SQL to clean and prepare a dataset on world layoffs for further analysis.
The dataset includes company details, industry, location, funding, and layoff counts.
The goal was to transform raw, messy data into a clean and reliable format.

---

Objectives

• Identify and remove duplicate records
• Standardize inconsistent values (company, industry, country)
• Convert dates into the correct format
• Handle missing or blank values
• Deliver a cleaned dataset ready for analysis or visualization

---

Cleaning Process

1. Remove Duplicates

• Created a staging table to protect raw data.
• Applied ROW_NUMBER() with PARTITION BY across key fields to detect duplicates.
• Deleted duplicate rows.

2. Standardize Data

• Trimmed whitespace from company names.
• Grouped similar industry names (e.g., all variations of Crypto → Crypto).
• Fixed country names (e.g., removed trailing periods in United States.).
• Converted date from text (MM/DD/YYYY) into proper DATE.

3. Handle Nulls & Blanks

• Replaced blank industry values with NULL.
• Filled missing industry values using self-joins on company and location.
• Removed rows with no layoff information (total_laid_off and percentage_laid_off both missing).

4. Finalize Table

• Dropped helper columns such as row_number.
• Ensured only clean, consistent fields remain.

---

SQL Concepts Applied

• Window functions (ROW_NUMBER() with OVER)
• Common Table Expressions (CTEs)
• String functions (TRIM(), TRAILING)
• Date functions (STR_TO_DATE())
• Joins and updates for data filling
• Conditional deletes

---

File Structure

• layoffs_cleaning.sql → Main SQL script with all cleaning steps
• README.md → Project documentation

---

Tools Used

• Database: MySQL
• Editor: MySQL Workbench

---

Next Steps

With the dataset cleaned, the next phase would be:

• Performing exploratory data analysis (EDA) to uncover layoff trends
• Building visual dashboards to present insights