Honors Theses and Capstones

Date of Award

Spring 2013

Project Type

Senior Honors Thesis

College or School



Mathematics and Statistics

Program or Major

Statistics & Mathematics

First Advisor

Ernst Linder

Second Advisor

Matthew Shump


A two stage modeling approach for modeling customer age as a predictor of contact volume was conducted using a real-world data set of approximately 2,000,000 contacts from a company call center. Two models were constructed in the first stage, one a straightforward regression and the other a series of regressions. One was selected as better performing and scaled up to predict calls received from calls answered. The second stage of the modeling included a day of the week covariate and performed the best of the models created. This model uses age bins as model effects, of which the youngest age bin had the largest effect. The model predicts the volume of calls received by the company.