Análises comparativas de vendas em um supermercado do município de Foz do Iguaçu por meio da mineração de regras de associação antes e durante a pandemia de COVID-19

Abstract

Discovered at the end of 2019, the Sars-CoV-19 has changed the way we live in Brazil, byclosing non-essential stores, schools, and with reduced hours for the public in supermarketsand pharmacies. In addition, mass layoffs occur across the country, which occur with theintention of reducing costs, thus directly altering the consumer’s income. In this sense, the population started to go less to the markets and buy only the essentials. With that, the interestarose in starting a research on the consumer shopping behavior, allied to the computationalfundamentals, such as the development of data capture algorithms in electronic invoices,through the Python programming language. Subsequently, it was intended to use the formattedand pre-processed data in a suite of artificial intelligence tools to perform data mining throughan association algorithm, the Apriori. With the need to be able to ascertain the behaviorsof several different fronts, it was determined that the entire survey would analyze three fullmonths of pandemic peak, these months, in turn, segmented and compared between them andalso separated by day of the week . In parallel, for work to be more reliable in terms of behaviorchange, for all bias analyzed during the pandemic period, periods without the consequences ofthe pandemic were also analyzed, that is, the previous year. After carrying out all the proposedanalyzes, the results of behavioral changes were considered interesting, after all, there werechanges in all periods, such as in the months of April and June, when the consumption of softdrinks, together with that of bakery products , soared, not only in comparison with 2019, butalso with the month of May, when this did not happen. It was also noticed that, unlike all otherdays of the week, only on Sunday there was a tendency to buy yogurt with bakery products, together with a third item that varied between crackers, cheeses, ground beef, among others.Finally, this research ended with the statement of reports to the specialist, who pointed out whatwas considered relevant and, with that, could help a supermarket to make decisions to try toincrease its sales.