🌍 Atmos Viz: Global Temperature Analysis and Prediction

Overview

Atmos Viz is a Python-based project designed to analyze, visualize, and predict global temperature trends across various cities and countries. Leveraging historical climate data, this project combines machine learning models, geospatial mapping, and interactive visualizations to uncover meaningful insights into temperature variations over time.

---

Project Features

• Data Preprocessing: Efficiently handles large-scale datasets, cleans missing values, and extracts features like Year, Month, and cyclic seasonal trends.
• Geospatial Mapping: Visualizes temperature data geographically using shapefiles and tools like GeoPandas and Plotly.
• Machine Learning Prediction: Forecasts future temperature trends using Linear Regression and Random Forest models.
• Interactive Visualizations: Generates engaging charts, heatmaps, and world maps for better analysis and understanding.
• Performance Evaluation: Measures model accuracy using metrics like MAE, RMSE, and R².

---

Technologies Used

• Programming Language: Python
• Libraries:
  • Data Handling: Pandas, NumPy
  • Visualization: Matplotlib, Seaborn, Plotly
  • Geospatial Mapping: GeoPandas
  • Machine Learning: Scikit-learn
  • GIS Tools: Shapefiles

---

Project Workflow

1. Data Preprocessing

   • Load the temperature dataset (sourced from Kaggle, originally from Berkeley Earth).
   • Clean and normalize missing temperature values using monthly means.
   • Extract cyclic seasonal trends using sine and cosine transformations.

2. Model Development

   • Use Linear Regression and Random Forest for predictive modeling.
   • Train models using features like Year, sin_month, and cos_month.
   • Evaluate performance using MAE, RMSE, and R² metrics.

3. Visualization

   • Geospatial Maps: Display temperature data using GeoPandas and Plotly.
   • Time-Series Trends: Plot temperature trends with trendlines for specific cities.
   • Interactive Maps: Visualize data with hover information for cities on the map.
   • Heatmaps & Boxplots: Analyze monthly and regional temperature variations.

Contributions

Contributions are welcome! If you'd like to improve this project, please fork the repository and create a pull request.

Contact

For any queries or suggestions, feel free to contact me:

Email: siddhantborse27@gmail.com

VPortfolio: Siddhant Borse

⭐ If you like this project, star it on GitHub! ⭐

---

Project Directory Structure

Atmos-Viz/
│
├── Package/
│   ├── data_preprocessing.py      # Data loading, cleaning, and feature extraction
│   ├── models.py                  # Machine learning model training
│   ├── evaluation.py              # Model evaluation metrics
│   ├── viz.py                     # Visualization functions (maps, charts, heatmaps)
│
├── shapefiles/                    # GIS shapefiles for mapping
│
├── data/                          # Raw and preprocessed temperature datasets
│
├── main.py                        # Main script for model execution and visualization
│
├── requirements.txt               # List of dependencies
│
└── README.md                      # Project documentation