Solar Odyssey is a data science project developed by students of the Batch 1142 of LeWagon Data Science Bootcamp in Berlin. The goal of the project is to predict the energy demand of rural areas in Nigeria based on satellite images using a model based on the densenet121 architecture.
This project used data provided by: http://rrep-nigeria.integration.org/#
There are a few known issues with the Solar Odyssey project that should be taken into consideration:
Model Accuracy: The current model's accuracy is not great, and further improvements are necessary to make more accurate predictions.
Atmospheric Reflection: Atmospheric reflection or haze above forested areas distorts images, which hampered the training and predictive power of the model.
Processing Power: Due to a lack of access to processing power, the entire densenet121 model could not be fine-tuned, which limits the potential accuracy of the model.
The Solar Odyssey project was created the following students of Batch 1142 of LeWagon Data Science Bootcamp in Berlin. The contributors to the project are:
- Karim Elbana
- Arnoud de Haan
- Josef Perara
- Jonathan Büning
The Solar Odyssey project is licensed under the MIT License. Feel free to use and modify the code as needed.