Airbnb is an online platform that enables property owners to put their vacancy room online so that guests can find an accomodation. In order to maximise the profit for the property owner and find a best deal for the guest, setting a price plays an important role in their retention in the platform. In this case study, I am going to use the dataset by Airbnb to predict the listing price based upon the feature.
(1) How does geographical location affect the price and How much does it contribute to the model?
(2) How does the room quality affect the price and How much does it contribute to the model?
(3) How does text data affect the listing price and how much does it contibute to the model
Outlier reduction (IQR measure)
Spatial Visualisation
Natural Language Processing
Feature Engineering based upon domain knowledge
Random forest with K-fold Random Search
Permutative future importance
RMSE