ISCAP Proceedings - 2023

Albuquerque NM, November 2023

2023 ISCAP Proceedings: Abstract Presentation

Applications Of Multidimensional Scaling With Co-plot Analysis

Priyanka Poosapati
University of North Carolina Wilmington

Yao Shi
University of North Carolina Wilmington

Judith Gebauer
University of North Carolina Wilmington

Yang Song
University of North Carolina Wilmington

Exploratory graphs play a crucial role in suggesting relevant statistical analyses and models while diagnosing various aspects of data during data analysis. Multidimensional scaling (MDS) is a widely used data analysis method that allows visualizing large and complex datasets in a smaller space, preserving pairwise differences or similarities between data points. However, to unlock the full potential of MDS, Co-Plot analysis has been introduced as a complementary data visualization technique. Co-Plot analysis extends MDS capabilities by presenting additional scatter plots or line plots alongside the traditional MDS plot, enabling a multidimensional view of relationships between variables from different datasets. This technique is particularly valuable when exploring with multidimensional data or comparing variables from different domains or measurement scales. Co-Plot Analysis aids in identifying patterns, clusters, correlations, and influential objects that affect the MDS plot's setup. In this study, Co-Plot Analysis is applied to: (1) multidimensional dataset of MBA Programs, focusing on the similarity among IT concentrations in MBA programs and the correlations among IT courses of different universities, (2) ecological community analysis, particularly for comparing species or communities across multiple sites or environmental contexts thus enabling the comparison of various groups within a dataset. The analysis is carried out using Non-Metric Multidimensional Scaling (NMDS) and Robust Multidimensional Scaling (RMDS) techniques. Shepard graph and Co-Plot Analysis is used to visualize the categorical variables in a reduced-dimensional representation, with points closer together indicating higher similarity. The goal in the study is how co-plotting works and the applications of multidimensional scaling where the Co-Plot Analysis can be used. The NMDS and RMDS techniques used here help us in understanding the Co-Plot Analysis better and provides enough flexibility for us to select the technique to produce better results. Keywords: Multidimensional scaling (MDS), Co-Plot, Robust Multidimensional Scaling (RMDS), Data visualization.