Contributed collaboratively by: Aishwarya Prashant Kamat, Qian(Lucy) Wu, Haridhakshini (Harisha) Subramoniapillai Ajeetha
Special thanks to my teammates for the great work! It is a pleasure to complete this project together! 💗
###################################################
Data Source: (Inside Airbnb) http://insideairbnb.com/get-the-data/
Code file: .ipynb document with comments.
This project involves preprocessing and data cleaning, exploratory data analysis, feature selection, model fitting and comparison, as well as recommendations and business insights.
The different algorithms applied and assessed for model-building include multiple linear regression, cross-validation, Lasso regression, CART (classification and regression tree), and Random Forest. In all models, it is found that using logged price (instead of price) as the response variable improves the prediction power significantly.
Based on two major performance metrics, we picked Random Forest Model as the final model for pricing suggestions. The result is beneficial for both the Airbnb website and potential hosts, as they can utilize this technique to rationally determine an appropriate price in the changing market conditions.