Extração de informações mercadológicas a partir de notas fiscais eletrônicas

Abstract

Most of the commercial transactions in Brazil are informed to the govenmet by eletronic invoice (NFe). In these invoices, there are information about the commercial transaction carried out. This information is avaliable to the general public throught the websites from the threasuty office in the Brazilians cities, with the right keys it is possible acces and mining data from the eletroninc invoices generated in Brazil, being able to have a huge volume of data containing information about all brazillian market segments. With the modelling and the generation of the database containing the data from the eletronical invoices it is possible apply techniques of data mining and extract information, includding georefenciated and in the time, about many aspects of the Brazilian market. One of the main problems addressed in this dissertation was the Pentup demand. From the information colleted in the eletronical invoices it is possible evaluate the existency of the Pent-up demmand about some product in a specific region. It is quantitatively evaluated, for exemple, a company open a branch in a specific region. For this analysis, some mecahnisms as: pattern recognition, clusterization and time series are employeed with the intent of consolidate the results. This dissertation presents a set of information about the comerce of many products in many fields of the brazillian market. It is expected have an relevant analysis about the dynamc of the laws that govern the Brazilian market, more specifically in the Metropolitan Region of Recife.