This project aims to investigate temperature changes over time and predict future temperature patterns on a regional and global scale. We employ time series forecasting methods, including neural networks, ARIMA, and SARIMAX, using the GISTEMP v4 dataset from NASA.

• Project Structure
• Data
• Analysis and Modeling
• Results
• How to Run
• Dependencies
• Contributing
• License

• global_temp_nn/: Jupyter notebooks for data exploration, cleaning, analysis, and neural network implementation.
• arima_sarima/: Jupyter notebooks for ARIMA and SARIMAX time series forecasting.
• images/: Contains images generated during the analysis.
• data/: Dataset files used in the project.
• README.md: Project overview and instructions.
• temperature_video_june.avi: Temperature anomalies plotted on a world map from 1880 to 2023.

The dataset used in this project includes:

• Global Surface Temperatures
• Northern Hemisphere Temperatures
• Southern Hemisphere Temperatures
• Zonal Temperatures
• Extended Reconstruction SSTs Version 5 (ERSSTv5) (NetCDF file)

• Exploratory Data Analysis (EDA): Initial exploration of the dataset to understand patterns and trends.
• Data Cleaning: Handling missing values, interpolation, and ensuring data integrity.
• Time Series Forecasting: Utilizing ARIMA and SARIMAX for time series forecasting.
• Neural Network Models: Implementing neural networks for more complex analyses.

• Visualizations of temperature anomalies over time.
• Forecasts of future temperature trends using ARIMA and SARIMAX.
• Neural network predictions for complex analysis of temperature patterns.

Download the zip file and run the individual Jupyter notebooks:

1. global_temp_nn.ipynb
2. arima_sarimax.ipynb

• TensorFlow
• Statsmodels

Install the required packages using:

pip install tensorflow statsmodels

Shubham Garg- [email protected]

Raunak Shukla - [email protected]

Phani Varma Gadiraju - [email protected]

This project is licensed under the MIT License.