Forecasting ATM transactions

Installing the ATMs to the right and effective points, which are the only channel providing cash transactions among Digital Channels, is the key point in terms of cost decrease and ease of customer access in the banking sector. The main purpose of this study is supporting installation of a new ATM decision both outdoor and indoor (in the Shopping Centers) by using data mining and forecasting techniques. We have used 1,115 existing ATM parks for Outdoor ATM transactions forecasting, and 195 points for Shopping Center ATMs Transaction Number Estimation. We have used data mining techniques because of the huge amounts of data. Multivariate regression analysis has been performed according to district / neighborhood population, other bank ATM clustering, number of customers in neighborhood detail, neighborhood saving and neighborhood household income, shopping center size, number of stores and parking capacity data. In this study, R-based R Studio program has been used for all classification and estimation methods.