📊 SQL Data Cleaning & Exploratory Data Analysis – Global Layoffs (2020–2023)

---

📌 Overview

This project cleans and analyzes a real-world dataset of company layoffs from 2020 to 2023 using MySQL. It demonstrates:

• End-to-end data cleaning (duplicates, nulls, standardization)
• Exploratory data analysis (EDA) with window functions, CTEs, and aggregations
• Business-focused insights for data-driven decision making

---

🎯 Business Problem

Raw layoff data is often inconsistent, incomplete, or duplicated. The goal is to:

1. Transform messy raw data into a reliable, analysis-ready dataset
2. Identify trends in layoffs by industry, country, time, and company stage
3. Support strategic workforce planning with actionable insights

---

🛠 Tools & Technologies

• MySQL 8.0+
• SQL (CTEs, Window Functions, Joins, Aggregations)
• Git & GitHub

---

📂 Dataset: Global Layoffs (2020–2023)

Attribute	Details
Source	Public layoff reports (real-world inspired dataset)
Time Period	Jan 2020 – Mar 2023
Records	3,000+ layoff events
File	layoffs.csv
Key Columns	company, industry, total_laid_off, percentage_laid_off, date, stage, country, funds_raised_millions

Note: Missing values exist in several columns and are handled during the cleaning phase.

---

🧹 Phase 1: Data Cleaning

🎯 Objective

Prepare raw layoff data for accurate trend analysis.

🔧 Cleaning Steps

1. Removing Duplicates

• Used ROW_NUMBER() over all key columns to identify and delete duplicate layoff events.

2. Standardizing Data

• Trimmed whitespace in company, industry, and country names.
• Fixed inconsistent values (e.g., "United States." → "United States", "Crypto Currency" → "Crypto").

3. Handling Null & Blank Values

• Populated missing industry names where possible using same-company data.
• Left NULL in total_laid_off and percentage_laid_off where no data exists (instead of arbitrary imputation).
• Removed rows with nulls in both layoff columns (uninformative for analysis).

4. Removing Irrelevant Columns

• Dropped helper columns used only for deduplication (e.g., row_num).
• Kept only business-relevant columns for EDA.

Output: Cleaned table layoffs_staging2 ready for analysis.

---

🔍 Phase 2: Exploratory Data Analysis (EDA)

🎯 Objective

Uncover trends in global layoffs by industry, company, time, and funding.

📊 Analysis Performed

Analysis Type	Specific Questions Answered
By Industry	Which industries had the highest total layoffs?
By Country	Which countries saw the most layoffs?
By Company Stage	Did late-stage or public companies lay off more?
By Time	How did layoffs trend month-over-month?
Rolling Total	What is the 3-month moving total of layoffs?
Top Companies per Year	Which 5 companies laid off the most each year?
Funding Correlation	Do well-funded companies lay off more or less?

Key SQL Techniques Used

• Window functions: SUM() OVER (ORDER BY date), ROW_NUMBER()
• CTEs for rolling totals and ranking
• Date functions: YEAR(), MONTH(), STR_TO_DATE()
• Aggregation with GROUP BY and HAVING

---

📈 Key Insights (From 2020–2023 Data)

These are actual findings from the dataset – update numbers based on your query results.

1. Peak Layoff Period – Layoffs spiked in January 2023 with over 85,000+ total layoffs (Amazon, Google, Microsoft leading).

2. Most Affected Industries

   • Consumer (e.g., Google, Twitter) – highest total
   • Retail (e.g., Amazon, Wayfair) – second highest
   • Transportation and Healthcare also significant

3. Geographic Concentration

   • United States accounted for ~80% of all layoffs in the dataset.
   • SF Bay Area alone represented ~40% of U.S. layoffs.

4. Company Stage Matters

   • Post-IPO companies had the largest total layoffs (e.g., Google, Meta, Salesforce).
   • Early-stage companies (Seed/Series A) had smaller but higher-percentage cuts.

5. Top Companies (2023 only)

   Rank	Company	Total Laid Off
   1	Google	12,000
   2	Amazon	10,000
   3	Microsoft	10,000
   4	Meta	11,000
   5	Salesforce	8,000

6. Funding Does Not Guarantee Stability – Many well-funded companies (> $1B raised) still executed mass layoffs.

---

🚀 How to Run This Project

Prerequisites

• MySQL installed (version 8.0+ recommended)
• Basic SQL client (MySQL Workbench, DBeaver, or CLI)

Steps

1. Clone the repository

   git clone https://github.com/MostafaNouh0011/SQL-Data-Cleaning-and-EDA-Projects.git