The goal of this project was to create a data-driven model to increase chances of success when flipping homes in Ames, Iowa. In order to achieve this, two research questions were accessed:
- How can we identify undervalued homes using data?
- What are the important levers for sale prices, and are any easier/cheaper to optimize?
Using data sourced from Kaggle's Ames Housing Dataset, several linear and tree-based models were accessed and the optimal was chosen based on accuracy and robustness. This model could then be used to identify under-valued homes and understand how to best increase their worth.
NOTE: Each model has it's own notebook to improve readibiliy and organization. Each model notebook will pull from the CSVs generated after running the Pre-Processing + EDA notebook.