ustcdj / airbnb_pricing Goto Github PK

Boston and Seattle Airbnb Data Analysis

• Introduction
• Installation
• File Descriptions
• Results
• Licensing, Authors, Acknowledgements

This analysis focuses on Airbnb listing price in Boston and Seattle and tries to answer questions below:

1. Is there any trend in Airbnb listing price? How does price compare between Boston and Seattle?
2. Should you become a superhost?
3. What factors affect the Airbnb listing price?

Please check my blog post for more details.

The link to the detailed codes: Airbnb.ipynb

The code was developed using the Anaconda distribution of Python, versions 3.8.1. Python libraries used are numpy, pandas, datetime, matplotlib, seaborn, sklearn, random, re.

The Seattle and Boston Airbnb datasets are from Kaggle. The Seattle dataset has 3818 records and the Boston dataset has 3585 records. The feat_info.csv file has necessary actions for each independent variable during the data preparation step.

• Preprocessing
• Drop features that have high correlations with other features.
• Drop records that have ridiculously high listing price.
• Drop rows with more than 30% missing values.
• Impute nulls with mean for numeric features and mode for integer/fixed values.
• OneHotEncoder categorical features.
• Scale features with StandardScaler.
• Split the data into train and test
• Use LassoCV to perform feature selection and to find best model (alpha) automatically

Within the same period (Sep 6, 2016 to Jan 2, 2017), daily average listing price is higher in Boston ($191) than in Seattle ($136).

Daily average listing price in both Seattle and Boston shows a strong seasonal pattern:

• relatively low at the beginning of the year
• gradually increases as the season enters summer
• highest in summer for Seattle
• highest in fall for Boston

Daily average listing price in both Seattle and Boston shows a strong weekday pattern:

• slightly high on Friday and Saturday, close to $6~$7 more than the average.

Compared to regular hosts, superhosts in both cities tend to:

• have on average 1 more review per month.
• charge extra_people.
• offer more services/amenities.
• are more pet friendly.
• Pets live on this property.
• require guest_profile_picture.

For both cities, top factors are:

• neighbourhood or location
• capacity, such as bedrooms, bathrooms, beds, accomodates, guest_included
• host_is_superhost, calendar_updated
• amenity — Doorman etc.

Please find the Licensing for the dataset at Kaggle Boston data and Seattle data. Other than that, feel free to play with the code here.