World Layoffs Data Analysis

Project Overview

This project analyzes global layoffs across various industries and companies. The data highlights the impact of layoffs on employees, companies, and industries. The main goals are to:

• Identify trends in layoffs across time, regions, and industries.
• Quantify the severity of layoffs and understand their correlation with factors such as funding stage.
• Provide actionable insights through visualizations and statistical analyses.

---

Dataset

File: layoffs_staging2.csv

The dataset contains 1,995 records with the following columns:

Column	Description
company	Name of the company.
location	Location of the company (city).
industry	Type of industry (e.g., Media, Retail, Technology).
total_laid_off	Total number of employees laid off.
percentage_laid_off	Percentage of employees laid off.
date	Date when the layoffs occurred.
stage	Funding stage of the company (e.g., Series B, Post-IPO).
country	Country where the company is based.
funds_raised_millions	Total funds raised by the company in millions of dollars.

---

Data Cleaning

The data cleaning process involved the following steps:

1. Removing Duplicates: A staging table was created, and duplicate records were removed to maintain data integrity.
2. Handling Null Values: Key columns such as total_laid_off and industry were carefully imputed or cleaned to avoid skewing analyses.
3. Standardization: Date formats and text columns were standardized for consistency.
4. Column Selection: Unnecessary columns were dropped for a cleaner dataset.

---

Exploratory Data Analysis (EDA)

The EDA process included the following:

1. Data Loading:

• The layoffs_staging2.csv cleaned dataset was used.

2. Layoffs By Industry:

   • Correlation between industry and total_laid_off.

3. Trends in Layoffs over Time:

   • Layoffs By Month.
   • Layoffs By Year.

4. Layoffs By Stage of Funding:

   • Correlation between stage and total_laid_off.

5. Countries with the Most Layoffs

   • Correlation between country and total_laid_off.

6. Additional Insights:

   • Companies with 100% layoffs were identified.
   • Top companies with the highest total layoffs were listed.

---

Key Insights

1. Industries Most Affected: The Consumer and Retail sectors experienced the highest layoffs, driven by economic downturns and shifts in business models.
2. Funding Stages: Companies at the Post-IPO stage of funding were more vulnerable to layoffs, showing that companies that go public are most vulnerable to layoffs.
3. Geographic Distribution: The United States and India were the most affected countries in terms of layoffs, reflecting their large startup ecosystems.
4. Trends Over Time: Layoffs peaked during periods of economic uncertainty and decreased as market conditions stabilized.

---

Recommendations

1. Support Vulnerable Sectors: Focus on providing support and funding to industries most affected by layoffs, especially during economic downturns.
2. Monitor Company Funding Stages: Monitor companies at specific funding stages for early signs of mass layoffs. Especially companies at the Post-IPO stage.
3. Geographic Focus: Policymakers should provide region-specific support, especially in countries with high layoff rates.
4. Improve Data Collection: More granular data on employee roles and layoff reasons can enhance predictive models for future workforce planning.

---

How to Use This Repository

1. Data Cleaning:

   • Refer to 01_Data_Cleaning.sql for SQL scripts used in data cleaning.
   • Ensure the cleaned dataset is ready for analysis.

2. EDA:

   • Refer to 02_EDA.sql and world_layoffs_eda.ipynb for exploratory data analysis.
   • Review the notebook as well as the world_layoffs_report.pdf for visualizations and insights.

3. Insights and Recommendations:

   • Summarized in this README file, Jupyter Notebook as well as the world_layoffs_report.pdf report.

---

Next Steps

1. Build predictive models to forecast future layoffs.
2. Incorporate additional datasets for a more comprehensive analysis.
3. Explore deeper causal relationships using advanced statistical methods.

---

Acknowledgments

• The dataset was curated from publicly available information on layoffs.
• Data cleaning and preparation scripts were authored to ensure data quality.