- Comparison of city districts
Slide #1: Introduction to AirBnB
-
Plot-map of all airbnb listings
-
Popularity developent over the years -- proxy by reviews
-
Plot interactions: -- filter by popularity, price -- slide by start-year and end-year
-
barchart of reviews per year -- interactions: onhover filtered by year
-- iframe: with singular listing
Slide #2: Hotels and AirBnB coverage
-
Plot of hotels in Copenhagen
-
Region area coverage for airbnb and hotels (week 8 / heatmap)
-
plot interations: -- checkbox for airbnb listings -- filter hotels and airbnb by popularity
-
region-plot interactions: -- change property from fixed values for density
Slide #3: Multi-Listing
-
Plot of airbnb listings
-
plot-interactions -- click/hover => highlight cluster on map (same host_id) -- dropdown-selection of top-multi
-
barchart of distributions -- slider to filter amount (default is 5 > ) -- todo: cool stuff with hover/click
Slide #4: Earnings AirBnb
-
Plot of airbnb listings
-
plot-interactions -- filter map by score -- click on listing => display tooltip (max earnings, popularity, approx earning, url) -- dropdown-selections of top-multi
-
Earnings per listing [interactive]
-
Earnings per host [interactive]
Slide #5: Reviews
-
Plot-map of all airbnb listings
-
plot-interaction: -- dropdown: Top # amount of reviews
-
bar-chart showing #bins = unique listing -- hover
Slide #6: AirBnB vs Hostels: Final Page
- Summarize with text!
Narrative: --- Multilisting guys => one bad review for a room in a hotel is very bad => one bad review for a room in airbnb not so much
=> AirBnb "hotel" is split throughout the city => Actual hotel has fixed location.
=> main point running a hotel with benifits (in comparrison to hotels of same size)
Consider if a hotel has a bad room and a customer complains, it affects the image of the entire hotel. But if a listing on Airbnb has bad review, it says nothing about other listings on Airbnb. Could that be considered unfair competition? Secondly, we know there's a guy who owns 256 listings on Airbnb. Now, most people don't know or care that this person rents out this many places. So if one place has a bad review, it doesn't affect the rating his other listings. Were it a hotel, it would. Obviously, this is an extreme example because this person with 256 listings is an outlier - but many hotels don't even have 20+ rooms, and there's a few select actors in the Airbnb business who do have 20+ listings on Airbnb.
That's a major difference; On Airbnb, the room itself is rated. With hotels, the entire hotel is rated. Equally, one well-rated room in a hotel is good for the entire hotel (so it goes both ways).
When we looked into reviews, we see that 94% of all reviewers have only done a single review of a listing in Copenhagen. The remaining 6% have done at least two reviews. We looked into the 6% and found that some people have left up to 20+ reviews. Looking into these reviewers, we find that they visit the same listings repeatedly. The top reviewer has visited the same listing more than 30 times, and also left 30 reviews - all which get shorter every time. Consider this; 30 good reviews from 30 different people has more credibility than 30 good reviews from a single person. So visiting the same hosting over and over cranks up a listing's overall score - and this is something Airbnb's system needs to take into accout.
Secondly: If visitors have very good relations with hosts, that means they can arrange bookings around Airbnb. A host may rent out up to 70 days maximum through Airbnb, but you can get around the limit when making booking deals outside Airbnb. Of course, we have no data so support that this happens - but we do have data that hosts can form year-long relations with their guests.
There is that matter that renting out through Airbnb is more lucrative than renting out to people staying long-term. (we have news as sources for this claim! We can look into earnings here.