Predicting Airbnb Licensure in Vancouver

Overview

This project aims to predict the legality of Airbnb's in Vancouver. Using Airbnb data collected from InsideAirbnb (data/yvr_listing_data.csv).

Through a multiple logistic regression model, we are able to predict the legality of an Airbnb with an accuracy of 0.90.

The repository has the following main jupyter notebooks:

1. data_cleaning.ipynb: This notebook is used to clean the data and prepare it for analysis. The cleaned data is saved as yvr_listing_data_cleaned.csv.

2. Logistic_Regression_Assumptions.ipynb: This notebook is used to check the assumptions of the logistic regression model. Modifications from checking the assumptions are saved to yvr_listing_data_cleaned_fixed.csv.

3. sm_model.ipynb & sklearn_model.ipynb: These notebooks contain the code for the models used to predict the legality of the Airbnb's.

4. EDA.ipynb: This notebook contains the code for the exploratory data analysis on the key variables discovered by our logistic regression model.

The repository also contains a src.py file which contains the functions used in the notebooks.

All data is stored in the 'data' folder.

Data Pipeline