Data Mining for Analysis and Prediction of Traffic Violations in the City of Recife

• Autor(a) principal: Torcate, Arianne Sarmento
• Data de Publicação: 2021
• Outros Autores: Barros, Maicon Herverton Lino Ferreira da Silva, Fonseca, Flávio Secco, Galindo, Marcos André Santos
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista de Engenharia e Pesquisa Aplicada
• Texto Completo: http://revistas.poli.br/index.php/repa/article/view/1679

The significant increase in traffic violations has become somewhat casual in the lives of Brazilians. The city of Recife, state of Pernambuco, according to the Dutch company TomTom Traffic, in 2018, ranked 10th among the cities with the worst car traffic in the world. In 2019 it ranked 15th. Given this, this research aims to investigate factors related to the increase in the amount of automatic measurement equipment and traffic agents. The objective is to create a model of prediction of violations in traffic by turns, tested with the real basis for the year 2019. Whereas the data selected for visualization and training of Machine Learning techniques were for the years 2017 and 2018, extracted from the open data portal. To guide the mining and data analysis process, the CRISP-DM methodology was applied. In addition, tools such as Pentaho PDI, Weka, GretL, Python and Orange Data Mining were also used to assist in this process. The results obtained indicate that there is a significant increase in infractions on holidays, mainly in Corpus Christi. In addition, monthly predictions show good results when compared to the actual numbers of infractions.