What are the most influential predictors of rental and sale price of an apartment? Part 2.

What are the most influential predictors of rental and sale price of an apartment? Part 2.

Posted by Ted Dogan

Updated: Nov 5, 2019

What are the most influential predictors of rental and sale price of an apartment? Part 2.

Understanding the complexity of economy requires detailed knowledge of leading and lagging indicators. This applies even with respect to how much does it cost to buy or rent an apartment in New York City. A tool that makes it easier to know what the market price would be comes in handy for that. With data scraping, it is possible to isolate the determinants of rental and sale price of an apartment in New York City.

For this project, I scraped Cityrealty’s website. This site provides rental and sale listings of apartments in NYC, in addition to providing history of sold apartments. About 20,000 observations related to intrinsic factors were collected for three categories: rental, sale, and sold listings. They consist of number of bedrooms and bathrooms, square footage, doorman, furnished, sub-boro, amenities, building type, building name, date, and price. My initial observation is that in apartment rental market, the highest rent is in a condo building followed by condop. With respect to apartment sales, though, townhouse building commands the highest price, followed by the condo. These figures do fluctuate, depending on the neighborhood and number of apartments available, but overall, there is a consistency patterns.

In the second stage of my intrinsic factors’ analysis, I made a visual variation comparison of rental versus sale market. It reveals that in the month of October 2019, rental market price fluctuates more than sale market. There are several reasons for the price fluctuation in the rental markets. Among them is the fact that investors treat those properties as an investment and prefer to rent them out while they hold on to their property in the expectation of property value increase over time. From an investor’s perspective, in this economic climate, especially considering the upward trend since the great recession, realizing the cash value of rentable properties might be an investment loss.

I also investigated sub-boro mean price comparison. It reveals that if a neighborhood is expensive when buying an apartment, it is also expensive when renting. This is also related to the income and education level of neighborhood, in addition to its safety ratio and business activity. Extrinsic factors are left out and analyzed separately.

The history of sold apartments follows a pattern similar to the neighborhood pattern. A detailed analysis of sold apartments data reveals that there is a bit more price fluctuation in January and March. As shown in the following graph, in January, condo and condop prices rise more compared to other building types. Since large volume of available apartments are in these categories, minor price fluctuation causes a market-wide rise. Price fluctuation in March can be related to a huge increase in townhouse prices. The graph shows that in March and April, townhouse prices rise disproportionally. I’m hoping to add more detailed demand analysis to understand these fluctuations.

Lastly, I created an OLS model to simplify many variables into a simple model. Out of 22 collected variables, only 9 are found to be statistically significant in explaining apartment rental, sale, and sold price. Interestingly, given the data and limitations, the building type turned out to be not a deterministic factor of apartment rental price. I searched for reasons and found that investors are buying properties in bulk, hoping that the value will keep increasing. This, in fact, supports my above argument about investors’ perspective. The model on intrinsic factors explain about 60% of variation.

I also looked at the literature and found many articles that identify which features add what amount to the cost of an apartment. In one of the articles, Stacy Sirmans says, she found that having a covered car park can increase the value of an apartment by about $50, with an additional increase of $15 for having a modern kitchen with a garbage disposal. She also found that accessibility to a public transportation has significant positive effects on rent. That was corroborated by another article written by Nishani. He stated accessibility to a public transportation has considerable impacts on rent. He also found that quality of construction, as well as age and the space of the building factor into the price. In the extrinsic part of this project, I found that a being one block distance to a subway station in NYC increases the rent by about $35. In addition, I defined a log-log model to better estimate the rent from external factors. The model includes only 5 variables out of 49 collected which explain almost 90% of variation and they are;

This is a multi-stage project. The first stage focused on the extrinsic factors and analysis of intrinsic factors is the the second stage of this project. The next step of the project will examine personal factors such as age, education level, single or married, have kids, have a pet, disability, race, personal income, smoking habits, and native country. In addition to understanding the impacts of those on rent, the project will include a research on the impact of Amazon delivery, Uber rides, and Airbnb rentals. I also intend to look into changing demographics, land value, and similar city comparison.

Each predictor that’s been collected thus far have some impact on rental and sale price. Some variables are statistically and economically significant, and some are neither. My goal is to create a model that explains the rent and sale price of an apartment, in addition to catching the changing dynamics of the housing market.

Ted Dogan

With the right approach and tools, it is possible to create a digital footprint of behavior. Actions leave traces of preferences, a raw form of demand. Converting them into a meaningful business story is an art that requires experience and intuition. And that’s where I come in. I am currently working on teaching machine “this feels like home.” In addition to having over a decade of work/entrepreneurial/teaching experience, I hold a master’s degree in finance and bachelor’s in economics. For more, I can be found at linkedin.com/in/teddogan

View all articles

Topics from this blog: NYC Machine Learning prediction Web Scraping python machine learning Student Works python web scraping

Interested in becoming a Data Scientist?

Get Customized Course Recommendations In Under a Minute