The discovery of patterns through data mining at IPTU in Curitiba

This article details the use of data mining on registration data and payment of the Urban Property and Land Tax (IPTU) in the city of Curitiba, for the months of January to June 2020. For this purpose, a database built in an Excel spreadsheet is examined, relating the payment of such tax to various attributes related to the characteristics of properties or unbuilt land. To carry out the analysis, an unsupervised learning algorithm for discovery of association rules (Apriori) in the Weka software (Waikato Environment for Knowledge Analysis) is used. The results indicate that by associating the different attributes of registration and payment, the payment of the tax is made in installments, regardless of the pattern of construction of the property or neighborhood. Another result obtained was that delinquency is higher in residential properties made of wood or masonry, with a finishing pattern: simple and simple average. Thus, given the results obtained, it is suggested that actions be taken by the municipal public administration to encourage the use of automatic debit and the payment of tax in cash by taxpayers, which would be very important for the cash flow and could minimize default in the payment of this tax.