Development of a Decision Support System in the Real Estate Market Using Statistical Data Analysis and Machine Learning Methods

Purpose of the study. The aim of the study is to develop a decision support system in the form of a Telegram bot, aimed at assessing the investment attractiveness of real estate objects using statistical data analysis and forecasting methods.

Materials and methods. The information base of the study is data from the platform – Central Real Estate Information Agency, containing information about residential real estate objects intended for sale and rent. The methodological base of the study includes methods of statistical data analysis, machine learning, as well as approaches to designing user interfaces in decision support systems. All necessary primary calculations and studies are performed using the functions of the Python programming language. The implementation of the decision support system was carried out in the Google Colab using the pyTelegramBotAPI library.

Results. Information was collected, cleared and pre-processed for 16 cities in Russia, and a study of rental and sale prices for housing was conducted. Using the CatBoostRegressor machine learning model, rental price forecasts for properties put up for sale were obtained, which made it possible to calculate their expected profitability. An analysis was also made of the possibility of using mortgage lending as a tool for increasing investment efficiency. A decision support system has been implemented in the form of a Telegram bot, capable of assessing the profitability of real estate and assisting the user in making decisions based on specified parameters and predictive models. The Telegram bot was tested, and examples of use were demonstrated, confirming the accuracy and usefulness of the calculations obtained.

Conclusion. The developed decision support system can provide recommendations based on the analysis of statistical data of the real estate market and a forecast model. The system is easy to use, focused on private investors, offers real objects presented on the market, automates the process of selection and evaluation of objects, and allows comparing purchase strategies using a mortgage and without attracting additional funds. 

1. Kovalevskaya N.YU. Methodological approaches to assessing the effectiveness of investments in residential real estate. Izvestiya Baykal’skogo gosudarstvennogo universiteta = Bulletin of the Baikal State University. 2020; 30; 1: 137-148. (In Russ.)

2. Shtark L.V., Sayenko I.A., Sharopatova A.V. Theoretical foundations for assessing the investment attractiveness of residential real estate. Russian Economic Bulletin = Russian Economic Bulletin. 2020; 3; 1: 84-87. (In Russ.)

3. Official website of CIAN [Internet]. Analiz rynka nedvizhimosti = Real estate market analysis. Available from: https://www.cian.ru/analiz-rynkanedvizhimosti-b2b/. (In Russ.)

4. Official website of Rosrielt [Internet]. Analitika rynka nedvizhimosti = Real estate market analytics. Available from: https://rosrealt.ru/analytics/. (In Russ.)

5. MakKinni U. Python for data analysis: data wrangling with pandas, NumPy, and IPython. O’Reilly Media; 2018. 541 p.

6. Zheron A. Hands-on machine learning with Scikit-Learn, Keras, and TensorFlow. O’Reilly Media; 2019. 1065 p.

7. Azibo: Real Estate Investing Software: Your Guide to the 10 Best Solutions.

8. PyTelegramBotAPI [Internet]. Ofitsial’naya dokumentatsiya biblioteki dlya raboty s Telegram Bot API = Official library documentation for working with the Telegram Bot API. Available from: https://pypi.org/project/pyTelegramBotAPI/.

9. Official GitHub website [Internet]. Sbor dannykh s sayta ob»yavleniy Tsian = Data collection from the Cian classifieds website. Available from: https://github.com/lenarsaitov/cianparseryu.

10. Official website of DOM.RF [Internet]. Stavki predlozheniy po ipoteke = Mortgage offer rates. Available from: https://nash.dom.rf/analitika/stavki_predlozheniy_po_ipoteke. (In Russ.)