Desenvolvimento de uma aplicação Web para avaliação automática de preços de imóveis habitacionais na Cidade de Maputo com base na técnica de Clustering

Abstract

The real estate market is a sector in which properties are traded, with the participation of owners, interested parties and brokers for sale and purchase. In Mozambique, traditional property valuation methods are widely used, based on historical data and the experience of appraisers. However, this approach can generate inaccurate results, especially in contexts where the evaluator has limited knowledge or comparative data is scarce. Furthermore, the subjectivity of the process can lead to inconsistencies in assessments. With the advancement of technology and the increasing need for faster and more consistent assessments, the use of machine learning was explored to modernize this process. An analysis of the current real estate valuation scenario in Maputo was carried out, trough which the comparative method was identified as the most used, however, this approach has shown limitations due to the lack of technological tools and price variations that do not reflect real market conditions. Therefore, it was proposed to develop a web application that applies machine learning techniques to estimate the value of properties based on historical data and relevant characteristics. The development of the solution included data preparation, feature engineering and model training, followed by the implementation of a prototype that allows estimates to be made based on information input by the user. During development, tools such as Django, Google Colab and MySQL were used to develop the application, analyze data and visualize results. The final solution offers a practical and objective alternative for evaluation, reducing dependence on traditional evaluators and providing users with a reliable reference for decision-making.