Project Information:
- Languages/Tools: R
- Project Date: March 2023 - May 2023
- GitHub: Click Here
Description:
Our analysis focused on predicting nightly rental prices for NYC Airbnb listings using a Random Forest model. Key preprocessing steps included handling duplicates and missing data, converting categorical variables, capping extreme prices at $500, and applying log transformation to stabilize variance. Feature importance analysis highlighted room_type, neighbourhood_group, and availability_365 as strong predictors of price variation.
Results:
The Random Forest model delivered the best performance, achieving an RMSE of 76.19 and explaining approximately 44% of the variance (R² = 0.437) in listing prices. This model is especially effective for providing price guidance in the $50–$250 range, making it a practical tool for new hosts and automated pricing systems. However, it is less effective at predicting prices for high-end or ultra-unique listings due to its averaging behavior.