Matheos Zaharopoulos
Navigate back to the homepage

Montréal Rental Market Analysis

Matheos Zaharopoulos
August 27th, 2020 · 5 min read

Montréal Rental Market Analysis - August 2020

On August 18th, I began my search to understand the renters’ market. With jobs becoming more remote, I wanted to know my renting options. I wanted to know the market trends for renting apartments, so my data mining operation is from a popular source. My friends warned me about scams, so I hope this helps me and maybe others avoid some. For this search, I checked 98 km from Metro Montmorency in Laval. This was a popular metro station with over 5 million passengers per year in 2019 (let’s be honest, this has changed) and a mega stadium. The large radius of search area covers enough ground to provide some options.

Highlights

  • Most rents are in Montréal
  • Most rents in August were from the end of the month*
  • Most rents were posted on Tuesday*
  • Personal exterior space increases rent costs
  • Pets significantly increase rent by nearly $200 per month
  • Included utilities have no effect on price and may be cheaper overall
  • Fridges are included in 31% of rents.
  • Appliances don’t add significant value (only 10% of units do not include any)
  • The rule of thumb: good deals fall within $750 and $1,600
  • Extra rooms cost between $147 up to $266
  • Rents are roughly $2.73 per square foot

*Note: August 18th is when the automatic mining operation began. There is no strict protocol about length, speed, and other confounds.


Database

My database holds over 4,000 listings, although not all results are relevant. This must do with the way they order results to cater to your search request. After exhausting results relevant to you, the database provides the next best thing. Naturally a robot will search through more pages than a human would.

Pie chart showing most rent postings are from Montreal

I filtered the results to cities with over 50 rentals saved to keep the results statistically significant. After filtering, four cities and 3,956 listings remained. The benefit is that many results ended up being from Montréal, probably because of the small 4-day mining period. It is also possible that Montréal has more units for rent and higher turnover of apartments.

A pie chart showing the reduced list of regions used to maintain adequate sampling criteria.

In the spirits of analysis, I wanted to see how dates harvested from the listings source code came into play for an apartment search. Moving dates seemed to be for September 1st when schools start. You can see most listings were posted towards the end of the month.

A horizontal bar chart showing August 18th having the most rents collected.

By day of the week, Tuesday was the most popular date. Keep in mind the note of the large confound in these last two graphics. The project started on August 18th (both high points) which shows the speed at which decision making can happen in real estate thanks to data science.

A graph that shows Tuesday as the day with the most frequent postings in august.

The Results

In this set, the first thing examined is the value of a personal exterior space. As one might expect, the more personal exterior space, the higher priced the rental on average. The error bars show that the obvious difference between no exterior personal space and a balcony may be as little as $50. The error bar (line the at top of the rectangle) shows the significant range of values around the average. A longer error bar shows that prices vary at wider ranges from the shown amount.

A horizontal bar chart showing the mean price of rents depending on exterior space.

Policies on Pets and Effect on Price

Our pets are important to us, so the next thing I checked was how that affects rental costs. The average rent for apartments that allow pets fully or limited is above those which do not. For those who are unsure how to read a box plot, the coloured portion shows the middle 50% of values with the line highlighting the average value. The lines at either end show the bottom and top 25% of values, with the dots being all those that fall above the normal distribution. Many dots in any direction often show that the average is less reliable of a measure. In our case, there is no significant difference gained from allowing pets, thus it is unimportant.

A boxplot showing the difference in rent depending on the landlords animal policies
A list showing the mean price of rents dependant on animal policies
An anova table showing significant difference in average price depending on animal policies.

The perspective of this article is macro considering a short drive around the city makes the differences between boroughs clear. Regardless, I was a little surprised that there is no mathematical difference between apartments with or without utilities included. Utilities have a minor effect on the price, perhaps indicative of a higher price attributed to ones without utilities. I suspect this has to do with a preference for independent heating for each apartment in new constructions.

Appliances and Utilities

A bar plot showing the average cost of appartments without utilities to be the same as those which include.

Appliances included seems like a regular occurrence. Of all the rentals, 31% came with a fridge and 24% with a laundry machine in the apartment. Only 10% of apartments had no appliances.

A pie chart showing the distribution of appliances included in rents.

Here is a bar plot indicating the value of each appliance. The trick and caution here are that it does not account for multiple appliances in the same price. This means, if most rents that include a dish washer also include laundry in the apartment then the high value of a dish washer is a lie.

A barchart showing rents that included dish washers are more expensive.

To avoid creating an extreme bias, I took another approach. The alternative approach considers every single combination listed and presents them in descending order according to the average price. It’s a bit of a mess and nothing mind blowing. Filtering the top 5 most popular combinations may have been more useful, although this version keeps more information. My summary is that fridge and freezer add nearly no value, neither does laundry in the building.

The effect of appliances in combinations on rent

In fact, you can see the average rent depending on appliance combinations does not have much logic nor does it jump by much between a few consecutive steps. We can see in the table below that rents without appliances average more in price than the first four appliance combinations.

The effect of appliances in combinations on rent averages

Price and Square Feet

The renters market has a wide range of prices and to avoid scams I drop some listings with extreme values. My interests in the next graph are rents between $750 and $2200 because this holds the top %50 of most common values. The most common rent is $2200 with 78 apartment listings. Another important range $1,000 (25% quantile) and $1,600 (75% quantile) because it represents a safe range of rents.

The distributions of prices in the database.

In terms of rooms, the prices all seem to be about what one would expect. The whiskers (top and bottom quantiles) overlap. The average rental prices seem to stagger by the number of rooms consistent with expectation. My feelings are that the prices do not vary a lot depending on rooms from one step to the next. For example, a 4 1/2 is $264 less than a 5 1/2 and a 5 1/2 is $264 less than a 6 1/2.

A graph showing the range of apartment square footage

The next table describes the difference in price from each step-up which ranges from $147 change to $266 if we neglect the “den step”.

Rents by number of rooms
The difference between rents at each step up.

Price and square footage have a strong correlation. This means that values cluster in a tight area well around a line and move in the same direction. Apartments cap off at 1,600 square feet, making the range of sizes enormous. The graph below shows a scatterplot with a linear regression plotted through it. A linear model suits this data and provides us with a formula to plug in square foot to find our price (or vice versa). Square footage of the apartment depends on the price in this graph to make the visuals clearer. The slope (or beta coefficient) is 0.367 which is the square footage you get for every dollar of rent. The opposite is also true, one square foot of apartment costs $2.73. This formula works best for apartments that fall within the normal distribution across the cities listed. It does not mean you should measure your apartment and do 400 * 2.73 = $1,090 rent. It wouldn’t be so far off if you lived in a trendy area that has a lot of action, but even then, the variation is wide by sub-divisions or boroughs.

A chart showing the correlation between apartment square footage and rent.
A scatter plot of price and square foot with a regression line.

A more granular perspective requires more data. This perspective is more robust at the expense of precision. Alternatively, it would be easy to parse the addresses or better yet, use Google. You can see an example of that in my article about private sales of homes in Quebec. If you’re interested in more specifics, please subscribe to my email list. I would be more than happy to respond to anyone’s questions through email at seo.department@outlook.com


Feedback is beyond welcome, it’s appreciated. I am testing these new template to provide articles with greater ease to focus on contact. Do you like it? Did you prefer the old one? Let me know.

More articles from Trends

How has COVID affected private sales of single homes in Québec?

With the Real Estate market in constant flux, how can you tell what's a good deal? Single homes are in high demand and low supply due to COVID-19. A detailed overview of the private sellers market.

August 15th, 2020 · 4 min read

Are Home Renovations Worth it in Québec?

When designing a home for a flip or resell value, it is worth knowing your choices. This article evaluates the added value from each design choice with a macro lense of Québec.

September 4th, 2020 · 5 min read
© 2020 Trends
Link to $https://www.linkedin.com/in/matheos-zaharopoulos-0906191aa/Link to $https://github.com/thematz1/