Predictive Analysis Using Web Scraping for the Real Estate Market in Gaziantep

Abstract

For investors and people who want to own a property, real estate is a crucial industry. Real estate includes land and any enduring construction, whether natural or artificial, such as houses, residences, apartments, and commercial structures. In Turkey, it is common to believe that owning property makes you live comfortably. Therefore, house ownership is a common aspiration among Turkish families. However, a variety of factors, such as a country's economic structure, inflation, world events, politics, etc., have an impact on the real estate market. In addition, the location, neighborhood, size, and number of rooms of a house can all affect how much it costs to live there. Gaziantep city is considered for analysis in the proposed study. The goal of this study is to predict which neighborhood, given a prospective buyer's financial status and specific property attributes, someone can afford to live in. As a result, web scraping is used to collect real estate data from the website. Once the data has been gathered, forecasting the neighborhood of a house is done using machine learning algorithms including decision trees, random forest, and extra trees. The results demonstrate that all algorithms produce good results with a performance accuracy of over 80%. However, among these algorithms, decision tree classification offers the best performance.