Airbnb-EDA-and-Regression

Airbnb-EDA-and-Regression

Colab · Kaggle · Report · Notebook · Video

Big data exploration and analysis on Airbnb dataset as well as regression model for price prediction of entities Airbnb Price Prediction Challenge

Problem definition

Airbnb, Inc. operates an online marketplace for lodging,primarily homestays for vacation rentals, and tourism activities. It is based in San Francisco,California. Our problem is going to be analysis and explorationon the given dataset we provided in the proposal and modeling the data in order to predictthe price of an instance.

Data overview

The dataset consists of 74,111 entries, splitted intothe following groups of data types.

Model selection and tuning and Results

The best outcome came for the XGBoost regressor , andgot a very good MSE with respect to
the published ones on the competition and near tosome of them and higher than some others
which is good as well as we have tried RandomForestRegressor and tuning done on ithad
very similar results but was a bit less than it aswell as trying Linear Regression but it was
less than them.

Training MSE: 0.2005
Validation MSE: 0.201

Training r2: 0.6092
Validation r2: 0.6139

Future work

• Use reviews dataset on each property and making sentiment analysis on them in order to add new feature to each property
• Using more larger dataset consists of multiple placesand countries

Contact

Mohamed Mokhtar - [email protected]