🌪️ Atlantic Hurricanes — Cleaning Walkthrough

This project demonstrates how I cleaned the messy Atlantic Hurricanes (Data Cleaning Challenge) dataset (1920–2020) and transformed it into a structured, query-ready SQL view.

---

📂 Files

• data/raw/Hurricanes.csv — original Kaggle CSV (CC0).
• sql/Hurricanes final clean.sql — end-to-end T-SQL script that cleans the raw data and builds the dbo.Hurricanes_Final view.
• data/clean/Hurricanes_Final.csv — optional export of the final cleaned dataset.

---

🔄 Reproduce

1. Create a database HurricanesDB and import data/raw/Hurricanes.csv into dbo.Hurricanes.

2. Run:

   sql/Hurricanes final clean.sql

---

This script will:

• Normalize messy text, whitespace, and quotes.
• Parse and standardize dates into StartDate and EndDate.
• Convert wind speeds into both mph and km/h.
• Normalize pressure into both hPa and inHg.
• Convert damage estimates (million/billion/k) into numeric USD.
• Derive DurationDays and Saffir-Simpson Category.

---

📊 Before vs. After

Raw data (sample)

Cleaned data (sample)

📝 Notes

• Dates are exported as YYYY-MM-DD.
• DurationDays = difference between StartDate and EndDate.
• Wind speeds and pressures are standardized for consistent analysis.
• Damage text like “Unknown”, “Minimal”, etc. are set to NULL.
• If opening the CSV in Excel, use Delimited → Comma import (not Fixed Width).

---

📖 Source

Valery Liamtsau, Atlantic Hurricanes (Data Cleaning Challenge), Kaggle, CC0 1.0.
https://www.kaggle.com/datasets/valerylia/atlantic-hurricanes-data-cleaning-challenge