☄️ Meteorite Landings Analysis (NASA Dataset) — Python + Power BI

This project explores over 45,000 meteorite landings collected by NASA, combining Python-based data science techniques with Power BI for interactive dashboarding. We deep-dive into mass trends, fall patterns, geographical distribution, and classification modeling to uncover insights hidden in cosmic debris.

---

🔗 Dataset Source

NASA Open Data Portal – Meteorite Landings

---

🛠️ Tools & Technologies Used

Tool	Purpose
Python	Data cleaning, visualization, modeling
Jupyter	Exploratory analysis & testing
Pandas	Data manipulation
Seaborn/Matplotlib	Visualization
Scikit-learn	Machine learning models
Power BI	Interactive dashboard & reporting

---

📊 Python Analysis

Performed in: metor.ipynb

✅ Key Steps:

• Data loading and cleaning (handling missing values, converting types)

• Exploratory Data Analysis (EDA)

  • Mass vs. Year scatter plot
  • Mass distribution by Fall type
  • Meteorite counts by decade
  • Pairplots to explore multivariate trends

• Hypothesis Testing

  • 1) does Meteorites that Fell Have Different Mass From that have Found ?

  • There is a difference in mass between 'Fell' and 'Found'.

  • 2)Has the average mass of meteorites changed over decades?

  • Significant difference in mass distribution across decades

• Feature engineering

  • Log transformation of mass for skew handling
  • Decade column for time analysis

• Machine Learning

  • Linear Regression: Predict meteorite mass based on features
  • Classification Model: Predict whether a meteorite was “Found” or “Fell” based on numeric and geographic features

---

📈 Power BI Dashboard

The Power BI report provides an interactive view of the meteorite landings dataset. It includes:

✅ Power BI Visuals:

• Map of global meteorite landings by mass and fall type
• Column chart: Meteorites per decade (with fall type breakdown)
• Box plot: Mass distribution by Fall type
• Bar chart: Most common meteorite classes (recclass)
• Histogram: Distribution of mass with binning
• Donut chart: Fell vs Found meteorite breakdown
• Line chart: Average mass over time
• Slicers: Interactive filters for class, fall type, decade, and mass range

---

📌 Key Insights & Conclusion

• Mass is extremely skewed — log transformation was key to uncover trends.
• Found meteorites tend to be heavier — possibly due to discovery bias.
• Meteorite discoveries have increased over the last century.
• Certain meteorite classes are far more common than others.
• Simple machine learning models showed promise in classifying fall types and estimating mass based on available features.

---

📁 Project Structure

├── Meteor/                            # Python analysis notebook and models
│   ├── Meteorite_Landing.csv    # Dataset folder
│   ├── meteor.ipynb                  # Main analysis notebook
│   ├── meteorite_mass_classify_model/  # Classification model files
│   └── meteorite_mass_linear_model/    # Linear regression model files
├── Meteor_Landings_Analysis.pbix     # Power BI dashboard file
├── README.md                         # Project documentation
├── LICENSE
├── requirements.txt                           # Project license file

🚀 How to Run

Python:

1. Clone the repo
2. Install dependencies: `pip install -r requirements.txt`
3. Run `metor.ipynb` in Jupyter Notebook

Power BI:

1. Open `powerbi-dashboard.pbix` in Power BI Desktop
2. Refresh data source if necessary

👨‍💻 Author

Your Name – Manas Sawant
Project for portfolio, learning, or data storytelling purposes.