Speaker
Description
The housing market is considered a critical sector and a driving force in nearly all economies. In Türkiye, ongoing economic instability has further amplified the significance of the housing sector, fueling a steady increase in housing prices nationwide. Given its critical role in the Turkish economy, the housing market’s trajectory is crucial not only for economic stability but also for household affordability. This study aims to predict housing market trends across various parts of the city by utilizing a comprehensive dataset of 20,000 rental listings scraped from a real estate platform. The dataset includes key housing attributes such as housing price, type, number of rooms, area, age, floor level, and geographic location. By integrating advanced data analysis and machine learning techniques, this research seeks to identify patterns and predict rental price movements, providing valuable insights for urban planning and housing policy development. Accurately predicting housing prices has been a critical issue that impacts a wide range of stakeholders, including landlords, investors, real estate agents, buyers, sellers, and institutions such as governments and banks.
Valuing housing presents a unique challenge due to its heterogeneous nature. This complexity has led most housing price studies to rely heavily on hedonic pricing theory. As Rosen (1974) explains, "Hedonic prices are defined as the implicit prices of attributes, derived from the observed prices of differentiated products and the specific quantities of characteristics associated with them." In simpler terms, the price of a house reflects the cumulative implicit values of its individual features, which cannot be priced separately. A hedonic equation is a regression model that links the price of a house to its various characteristics. In previous studies, the hedonic price model based on linear regression, which uses OLS estimation, has been often used to predict housing prices. However, the relationship between property value and its influencing factors is often complex and non-linear. Various other techniques have been developed that offer potential advantages over the standard approach. Recent research has demonstrated that non-linear models leveraging machine learning techniques provide more accurate and reliable predictions. These modeling techniques can be more accurate than the standard approach because they can learn from data and continually enhance their predictive performance. Artificial neural networks, gradient boosting, and random forests are the most fundamental and widely used machine learning algorithms for prediction. As a result of this study, it has been seen that machine learning methods make better predictions in line with the literature.
The second phase of the study focuses on evaluating the affordability of housing in different parts of the city, with significant implications for urban planning and housing policy. Housing affordability can be measured using various methods, such as the price-to-income ratio (PIR), which is defined as the ratio of the median house price to the median household income in a city or country. Using the price-to-income ratio, the study identifies areas at risk of affordability crises, gentrification, or housing shortages. For instance, neighborhoods experiencing rapidly rising rental prices can be flagged for interventions like affordable housing initiatives or rent subsidies. Conversely, areas with stagnant or declining prices can be targeted for revitalization efforts to prevent marginalization.
In conclusion, this research offers a robust framework for predicting housing market trends, combining machine learning, spatial analysis, and urban policy considerations. By providing actionable insights into rental price dynamics, it contributes to more equitable and sustainable urban development, addressing the housing needs of diverse populations in Ankara and beyond.
References
Rosen, S. (1974) Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition. Journal of Political Economy, 82, 34-55.
| Keywords | Housing Price Prediction; Housing Affordability; Machine Learning; Türkiye; Ankara |
|---|---|
| Best Congress Paper Award | Yes |
