Layoffs SQL Project — Data Cleaning & Exploratory Data Analysis

This project showcases my end‑to‑end process of preparing a real layoffs dataset for analysis using MySQL. The goal was to approach the data the same way I would in a professional environment—clean it thoroughly, validate assumptions, explore trends, and document clear insights.

---

Project Structure

sql-portfolio/
│
├── data/
│   └── layoffs.csv
│
├── sql/
│   └── layoffs_data_cleaning.sql
│
├── eda/
│   ├── layoffs_eda.sql
│   └── screenshots/
│       ├── summary_stats_1.png
│       ├── summary_stats_2.png
│       ├── top_companies.png
│       ├── yearly_layoffs.png
│       └── monthly_trends.png
│
└── README.md

---

1. Data Cleaning (MySQL)

All cleaning logic is in layoffs_data_cleaning.sql.
The raw dataset contained duplicate entries, inconsistent category values, missing fields, and unformatted dates.
My cleaning steps included:

• Creating Working Copies

I created layoffs_copy1 and layoffs_copy2 to preserve raw data and ensure each stage of the cleaning pipeline was traceable.

• Removing Duplicates

Used ROW_NUMBER() to identify and remove exact duplicates across all key columns.

• Standardizing Date Formats

Converted text‑based dates into proper DATE format using STR_TO_DATE, enabling reliable time‑based analysis.

• Normalizing Text Fields

• Trimmed whitespace
• Cleaned inconsistent industry labels
• Removed trailing punctuation

• Handling Missing Values

Where possible, missing industries were filled using information from other rows of the same company.

• Final Output Table

All cleaning steps result in a fully standardized and analysis‑ready table: layoffs_copy2.

---

2. Exploratory Data Analysis (EDA)

All EDA queries are stored in layoffs_eda.sql.
The analysis focuses on identifying trends across companies, industries, years, and months.

Summary Statistics

• Maximum layoffs and highest layoff percentage
  (see summary_stats_1.png)
• Date range of the dataset
  (see summary_stats_2.png)

Companies with the Highest Layoffs

Grouped layoffs by company to identify major contributors.
(see top_companies.png)

Yearly Trends

Analyzed layoffs across years to understand broader industry changes.
(see yearly_layoffs.png)

Monthly Trends and Rolling Totals

Explored month‑to‑month changes and cumulative layoff patterns.
(see monthly_trends.png)

---

Key Insights

• Layoff activity spikes significantly during 2022–2023.
• The largest tech companies account for a major share of total layoffs.
• Some early‑stage companies show 100% workforce reduction, indicating shutdowns or full restructuring.
• Dataset starts in early 2020, aligning with global economic disruptions.

---

How to Run the Project

1. Import layoffs.csv into MySQL.
2. Run all steps in sql/layoffs_data_cleaning.sql.
3. Query the cleaned dataset:

SELECT * FROM layoffs_copy2;

4. Run EDA queries from:

SOURCE eda/layoffs_eda.sql;

---

Skills Demonstrated

• SQL data cleaning and transformation
• Window functions for deduplication and rolling analysis
• Date conversion and validation
• Trend and aggregate analysis
• Structuring a clean, reproducible SQL workflow
• Presenting results clearly with supporting visuals