-
The first model creates a pricing tool for a first Airbnb property using features just related to the property, the location, and amenities. The predicted price will be used as a feature in the second model.
-
The second model will be set to maximise the user repose on a listing and predict an overall consumer review. The target will be defined in the process.
For feature selection, the Genetic Algorithm is used.
The model interpretability will be discussed using SHAP SHAP (SHapley Additive exPlanation) and how to leverage NLP for modelling and analysing the neighbourhood description given the overall score review. The folders contain:
The folders contain:
-
Code: All the code for building the two models requires (Read README in the folder)
-
Data: Raw data.
-
sample_data_dictionary_02: Draft of data dictionary and tabs used during the feature selection.
-
Presentation_Listings_ABR: slide with technical results