Airbnb is one of the most popular global online marketplaces to provide hospitality services to both guests and hosts. It not only provides opportunities for hosts to determine price and other details for their listings, but also offer guests a place to rate and write comments for each listing.
Give suggestions and predictions on price to Airbnb hosts given following scenarios:
Group A: People considering about getting a new house/apt and put it on Airbnb
Group B: People having specific listings already
Source: http://insideairbnb.com/get-the-data.html
Data Size: 49748; Number of Variables:106
Remove rows with NaN, get dummy variables on categorical features, and add new features
Give suggestions on purchasing a new property in NYC
Perform linear regression and random forest on features related to price and rating
Select important features to consider for improvements
Sentimental analysis and word cloud on 4 price&rating groups
Give suggestions on pricing and rating to be a Superhost
-
Location: in midtown or lower town Manhattan or brooklyn area that is near to Manhattan
-
Property types: loft or serviced apartment
-
Unit types: 2b2b, 3b2b, etc. (The ratio of number of bathroom to number of bedroom should be greater than or equal to ½)
For people in Group B, we can give them suggested prices of their listings based on our models before and after having rating data.
If they want to get a higher price and rating, they have to be careful about accommodates, number of available days, distance to Time Square and whether it's entire house or not.
Also, they can include “beautiful”, “neighborhood”, “home”, “spacious”, “great location”, etc. in summary to make their listings more attractive.