Towards Hierarchical Cluster Analysis Heatmaps as Visual Data Analysis of Entire Student Cohort Longitudinal Trajectories and Outcomes from Grade 9 through College

• Published in: The High School journal Vol. 106; no. 1; pp. 5 - 36
• Main Authors: Bowers, Alex J., Zhao, Yihan, Ho, Eric
• Format: Journal Article
• Language: English
• Published: Chapel Hill University of North Carolina Press 22.09.2022; The University of North Carolina Press
• Subjects: Academic Achievement; Big Data; Cluster analysis; College admissions; College Students; Data analysis; Data mining; Data science; Education; Educational research; Grade 9; High school dropouts; High School Students; High schools; Information Systems; Learning Analytics; Longitudinal Studies; Multivariate Analysis; Outcomes of Education; Prediction; Schools; Standardized Tests; Student Needs; Student Records; Students; Success; Teachers; Visual Aids; United States
• ISSN: 0018-1498; 1534-5157; 1534-5157
• DOI: 10.1353/hsj.2022.a906700

Research on data use and school Early Warning Systems (EWS) notes a central practice of researchers and practitioners is to search for patterns in student data to predict outcomes so schools can support success when students experience challenges. Yet, the domain lacks a means to visualize the rich longitudinal data that schools collect. Here, we use visual data analytic hierarchical cluster analysis (HCA) heatmaps to pattern and visualize entire longitudinal grading histories of a national sample of n=14,290 students from grade 9 to college in every enrolled subject and year, visualizing 6,728,920 individual datapoints. We provide both the open access code in R and an open-access online tool allowing anyone to upload their data and create a HCA heatmap, providing support for visual data analytic and data science practice for both education researchers and schooling organizations.