Applied Computer Vision on aerial images of urban areas to extract metrics of sustainability in the area, available for free, over many regions and over time. For this project Google Earth images were used for object detection and semantic segmentation. The models were trained on Google Colab to identify trees (as representative object) and segment the surface into four types. This proof of concept shows the informational value and potential of this data.  

The project is deployed with Streamlit and hosted on GitHub.