Python Code: Satellite Image Classification Project

Link to Kaggle dataset: https://www.kaggle.com/crawford/deepsat-sat4

To run the data processing, model training, and model evaluation Python notebook:

• recommend a minimum of 64 GB memory and 16 CPU cores (the specs of the Google Cloud VM that I used)
• dataset .csv files need to be uploaded to the directory where the notebook is being executed
  • X_test_sat4.csv
  • X_train_sat4.csv
  • y_test_sat4.csv
  • y_train_sat4.csv

To run the visualizations Python notebook, for showing predictions and images:

• Can be run in a free Google Colab instance
• Upload predictions file
  • to Colab working directory
  • or to Google Drive and use google.colab.drive.mount (as implemented in the notebook)
• Predictions file link

---

Back