🦈 Shark Attack Data Analysis

From Raw Chaos to Visual Intelligence

---

🌊 Project Story

Data rarely arrives clean. It crashes in like waves… messy, unpredictable, and full of hidden signals. This project dives into a Shark Attack Dataset and transforms it into meaningful insights using a blend of data cleaning, exploration, and advanced visualization.

Think of it as turning scattered ocean noise into a readable map of patterns.

---

🎯 What This Project Does

🔍 Investigates real-world shark attack data 🧹 Cleans and structures messy information 📊 Compares before vs after cleaning visuals 🚀 Uses advanced interactive visualizations to uncover hidden patterns

---

📂 Dataset Snapshot

• Dataset: Shark Attack Records

• Customization: Random 200-row subset (ensures uniqueness)

• Key Attributes:

  • 📅 Year
  • 🌍 Country
  • 🏄 Activity
  • 🎂 Age
  • 👤 Sex
  • ⚰️ Fatal (Y/N)

---

🧹 Data Transformation Journey

Raw data → Clean data → Insightful patterns

Key Steps:

• Removed irrelevant & noisy columns

• Handled missing and inconsistent values

• Converted data types (Age, Year)

• Eliminated duplicates

• Engineered new feature:

  • ⚡ severity (based on fatality)

---

📉 Before Cleaning

(The messy ocean 🌪️)

• Missing values visualization
• Unstructured country distribution

👉 Data looks inconsistent and hard to interpret

---

📈 After Cleaning

(Clarity emerges ✨)

• 📅 Year-wise attack trends
• ⚰️ Fatal vs Non-Fatal comparison
• 🎂 Age distribution

👉 Patterns become visible and meaningful

---

🚀 Advanced Visualizations

(Where the project truly shines 💎)

• 🌍 Choropleth Map → Global distribution of shark attacks

• 🌳 Sunburst Chart → Country → Activity → Fatality hierarchy

• 🎻 Violin Plot → Age distribution with density insight

• 🧬 3D Scatter Plot → Multi-dimensional data exploration

• 📦 Treemap → Compact hierarchical visualization

---

🛠️ Tech Stack

Tool	Purpose
Python	Core programming
Pandas	Data manipulation
Matplotlib	Basic visualization
Plotly	Advanced interactive charts

---

📊 Key Insights

• 🌍 Certain regions show higher attack concentration
• 🏄 Activities like surfing/swimming increase risk
• ⚰️ Fatal attacks are relatively rare but critical
• 🎂 Specific age groups appear more vulnerable

---

▶️ Run the Project

1️⃣ Install dependencies

pip install pandas matplotlib plotly

2️⃣ Launch notebook

jupyter notebook

---

📸 Visual Highlights

🌍 Global Distribution

🌳 Sunburst Analysis

🎻 Age Distribution

---

🧠 What Makes This Project Stand Out

✔️ Uses a custom dataset subset (not repeated) ✔️ Shows before vs after cleaning comparison ✔️ Combines basic + advanced visualization ✔️ Includes feature engineering (severity) ✔️ Focuses on storytelling through data

---

📌 Final Thought

Data is not just numbers. It’s a story waiting to be revealed.

This project demonstrates how thoughtful preprocessing and powerful visualizations can turn raw data into something meaningful, insightful, and impactful.

---

👨‍💻 Author

Bharath JR B.Tech CSE (AI & ML)

---