- Data Engineering: Merging data of different shapes (land prices, buildings, Zensus, etc.)
- Generating features: Building new features by exploring other data streams such as OSM (number/length of travel paths) and World Cover (proportion of grassland, etc.)
- Feature reduction: Using PCA we projected a total of 200 features to 20 while maintaining 95% of the data science variance
- Model evaluation: Examination of different models (see Model Selection file) and scoring analysis
Locally
- Install Python and Anaconda
- Run following command to create the anaconda environment with all necessary dependencies:
conda create --name tarantulas --file requirements.txt - Activate environment with:
conda activate tarantulas - Run the application by:
streamlit run app.py
On the Remote Server
Access the app by clicking This App Link
- Rohan Walia
- Yichen Zhang
- Lucas Vicentim
- Billy Herrmann
- Sandro Barrios