United States Bill Payment Machine Location Modeling
Jiawen Liu, Graduate Student, Oklahoma State University
Paying bills without a bank account can be challenging. The company sponsoring this project is a provider of kiosks which facilitates customers to pay their bills via cash, checks, and credit cards through an automatic process. The bills that can be paid include utilities, phone and insurance. This project is aimed to build a predictive model to estimate the total number of payment transactions in a retail stores on an annual basis. JMP® Pro 10 is used to build predictive models; more precisely data preparation uses JMP® Scripting Language. The dataset provided by the sponsoring company contains information about kiosks located in USA. Final dataset used in this poster contains information about 78 kiosks only. In this dataset, market related variables such as demographics, direct bills, and share influence variables, such as store type, brand, and competitors all collected within 1 mile, 3 miles, and 5 miles radius of each kiosk. The data is then standardized and transformed for later analysis. Regression and Factor Analysis followed by factor-bases scales are used in this project. As per our NDA with the sponsoring company, all transaction related numbers have been masked.