Datová analýza českého realitního trhu

The goal of my undergraduate thesis is to apply web scraping techniques and machine learning methods to the analysis of the Czech real estate market. This thesis focuses on data extraction using web scraping from a real estate portal. The data collected is intended to reflect the current state of the Czech real estate market. Furthermore, the thesis deals with data analysis using Python libraries and the results of the analysis are visualized in Jupyter notebook. Finally, the results are compared with the data from last year and the evolution of the data over time is described. The thesis is divided into theoretical and practical parts. The theoretical part introduces the field of web scraping, its principles, existing solutions and ethical aspect. Furthermore, the thesis describes data cleansing and the possibilities of storing and working with data. Popular data mining and data analysis methodologies are also described, with emphasis on CRISP-DM methodology. The last section is devoted to visualization theory. The practical part of the thesis starts with data collection from a web-based real estate portal www.bezrealitky.cz, including the development of a Python script. This is followed by an introduction of the tools used during the analysis and an introduction to the domain of real estate data. The obtained dataset is then preprocessed in the Jupyter Notebook environment. The resulting data is first analyzed using exploratory analysis and then using modeled with classification, regression and clustering data mining methods. Finally, the results are compared with the previous year's data, the evolution of the data over time is described, and the entire paper is summarized.