Show HN: Using Machine Learning to Find Undervalued Super Bowl Rentals(pricemethod.com)
pricemethod.com
Show HN: Using Machine Learning to Find Undervalued Super Bowl Rentals
http://www.pricemethod.com/superbowl/guest
5 comments
How are you able to use the term Super Bowl without getting sued?
They're not using it on their product name, it's merely a search keyword, why'd they get sued?
Where does Machine Learning play into this? For the estimate?
In that case, what's the training set?
In that case, what's the training set?
We use a linear model with binary extension of several properties of a listing to learn the current value of different listings as it is on the market.
We combine this daily with an analysis of how much hosts and hotels are already increasing their prices and whether they book or not to see how much our learned base price should be increased for the Super Bowl.
The dataset is based on publicly available data from Airbnb and Homeaway.
We combine this daily with an analysis of how much hosts and hotels are already increasing their prices and whether they book or not to see how much our learned base price should be increased for the Super Bowl.
The dataset is based on publicly available data from Airbnb and Homeaway.
"We use a linear model with . . . . . " ... that's what most HN'ers clicked the link for; and not to actually search listings. I wish you had linked to a nice, long nerdy writeup about the product, instead of the product itself (which, while interesting in general, is not as much for this crowd).
Great call on sharing a more robust write-up. We're going to be watching the market closely over the next 10 days and will share more here.
For our team, one of the more interesting aspects of continued observation will be a more fine-grained understanding of price elasticity around such a major event.
For our team, one of the more interesting aspects of continued observation will be a more fine-grained understanding of price elasticity around such a major event.
May just be me, but none of the links actually lead to the relevant airbnb listing. They all revert to area search.
If the link leads you to an ‘area search’, that means the user removed their listing from Airbnb, likely after having booked their place for the Super Bowl. Often, these are people who _only_ listed their place for the Super Bowl, a trend we expect will continue around other big events.
We analyze the market throughout the day, and on each pass, remove these listings from our results page. In the mean time, those ‘area searches’ illustrate the speed at which vacation rentals are booking for the Super Bowl!
- I'm Andrew, co-founder of PriceMethod. Thanks for the comment!
We analyze the market throughout the day, and on each pass, remove these listings from our results page. In the mean time, those ‘area searches’ illustrate the speed at which vacation rentals are booking for the Super Bowl!
- I'm Andrew, co-founder of PriceMethod. Thanks for the comment!
I'm really interested in the methods used to create this service and the learnings about rental behaviors associated. I'm also interested how this model might be able to scale to other cities and times to provide a continuously evolving model of price trends and predictions for goods and services.
The website is more or less algorithmic "magic" wrapped in a pretty UI, and I was hoping that you might have a blog post linked somewhere on the site so that I could read more about it.
The website is more or less algorithmic "magic" wrapped in a pretty UI, and I was hoping that you might have a blog post linked somewhere on the site so that I could read more about it.
Yep - agreed. We're a relatively new team, so there is a lot more to come. Our larger goal is around making a lot of data accessible and actionable for owners everywhere. Stay tuned!
You should take a look at this API: https://www.mashape.com/zilyo/zilyo
<snark>Just what superb owl attendees are most keen to do: save money! An amazing coup, a triumph really!
</snark>
How about just finding cheap apartments (or real estate) generally instead? Or is there too much manual intervention required in this machine learning?
I guess it's easier to track fluctuation of real estate value in Glendale, AZ rather than NYC or SF. I don't know OP though so I can't be 100% sure.