Vancouver Room Prices
Have you ever wondered what is a good price for a room in Vancouver?
As part of a data project I’ve been collecting data and analyzing the room rental market in Vancouver. Having moved several times in the past few years I thought that would be an interesting topic.
This post is a easy to read summary for everyone to read. I’ll leave the discussion of the methodology and a more advanced analysis in a set of future posts for the interested, nerdy reader:
- Part 1: Scraping Websites with Python and Scrapy
- Part 2: Data Cleaning and Geolocation with Python and Shapely
- Part 3: Natural Language Modeling and Feature Selection in Python
The data
Data was collected from a popular website for room listings at approximately 1000 posts/day, collected every night at 11:30. The information extracted includes price, geographical coordinates, title, content and other attributes of each post.
In the current post we examine data collected from March 31st to April 15th
Average room prices by neighbourhood
As you may expect, the neighbourhood significantly correlates with price. The following map shows the average post price by neighbourhood, greener areas are higher prices.
So far Downtown wins as the neighbourhood with the higher average price per listing, followed closely by Fairview. The west is generally on the expensive side, while the east is quite a bit cheaper.
Keep in mind that the numbers are just average prices. This doesn’t mean that a room in Downtown will generally cost more than an equivalent room in Victoria-Fraserview. For example, it may just be that Downtown prices are higher because most of the listings are luxurious apartments while Victoria-Fraserview is a suburban area where you rent basements and other humble rooms.
What do people say and what do people charge
By analyzing the title and the content of each posting, it is possible to determine how the presence of certain words (what they say) correlates with price (what they charge).
In the following chart you can see that when the post contains homestay, the price is an average extra ~$110 while basements are a bit less (~$30) than the average.
Interestingly, if the word female pops up, the price is also slightly lower, good for you girls!
You can also notice an issue in the model: the sublet keyword correlates with a higher price, contrary to what is commonly observed. This is very likely a drawback of the methodology (i.e. this is not a randomized experiment). One reason for this correlation could be, for example, that sublets are usually better rooms with a higher baseline price.
How long does a post stay up?
According to the data, about 2/3 of the posts are gone within 2-3 days, after that they follow a steady decline. Nevertheless, posts don’t hang around for too long, so write to a lot of people and quickly.
I hope you liked the insights, please write any comments/sugestions or, or if you wish, ask questions.
