Application of association rule method in data mining on customers using credit cards and reporting in power bi

Today, the number of customers using credit cards in the banking sector is increasing. The reason for this is that people need this way in order to adapt to changing life conditions and demands. It is getting harder and harder to analyze the customers due to the increasing number of customers. One of the most important factors that banks focus on after issuing a credit card is the determination of credit card payment terms. It is the analysis of customer qualities according to their credit status. In order to analyze the customer among the large data piles here, using data mining method is to reveal the relations and rules between the raw data piles. In this study, it is to determine the characteristics of customers according to credit card payment terms, thanks to the association rules method, which is one of the data mining methods. In the study, a monthly dataset covering customer information of the banking sector was used. Data analysis was performed on the dataset using the SQL language, and the information of 25,134 customers was made ready for the algorithm. In the data set, the customer's gender, whether or not they have a car, whether they have a property or not, the number of children, the amount of annual income, education level, marital status, age status and, accordingly, credit card payment terms were evaluated. However, it was used by categorizing it in certain categories for the areas determined in the data set. In the study, customer information was converted into a logical structure for the association rule, and support and trust thresholds were determined separately for credit card payment terms. For the solution of the study, association rules were created with the help of Python by using the apriori algorithm. In addition, the data is visualized thanks to Power BI. In this way, the characteristics of customers according to credit card payment terms were determined. As a result of the study; Adult customers who do not have a credit card, pay their credit card on time, and spend 3 months past the credit card's maturity date are female. Adult customers who have no children are those who have overdue their credit card debt and have bad debt. In addition, rules such as those who do not own property and who have graduated from primary/secondary education are customers who do not pay their credit card debt even after 6 months have passed. As a result of these results, it aims to determine the credit card payment terms of current and potential customers. Keywords: Data mining, association rules, apriori algorithm, banking industry, Power BI, Python, SQL