486-487 / 568
Recommendation System for Airbnb
Background:
Let’s say you found an Airbnb location that you liked, and you decide that it is too far
from the nearest subway station. You wish to find similar listing somewhere close to
station. by choosing another member of the same cluster closer to the station, you
can trust that there is some similarity between these two locations. You can then,
narrow down your search and find what you are looking for.
Rather than browsing through hundreds of Airbnb listings by reading their
descriptions, checking their amenities, filtering by price, or asking your friend’s
opinion, you can easily find similar Airbnb listings that meet your requirements!
Methods & Process:
Nadav Kiani, Daniel Dimant, Benny Barki
Advisor: Ph.D. Or Givan
Industrial Engineering
System Integration:
Users can enter our website and conduct a search as shown below:
Our website is integrated with the dataset, so once the query is executed, the users
will be directed to a web page which displays a dashboard that contains information
about the filtered listings in the specified neighborhood.
Our system can help Airbnb users make a better
decision when they want to book a listing. Our system
finds groups of similar listings based on the listings
features. This helps minimize the number of listings
the user has to review in order to find the best for him
by looking only at a specific group instead of the entire
search results. Our system produces an interactive
dashboard that contains information about the listings
in a specific neighborhood in New York City, that can
help the user understand how each group differs from
one another.
Data Cleaning
& Feature
Engineering
Scaling
Dimensionality
Reduction
(PCA)
Clustering
(K-Means)
Build our web
framework
(Flask)