Exploratory data analysis of dental health variables in Irish teenagers

Objectives: To carry out exploratory data analysis Of longitudinal datasets highlighting the importance of appropriate data visualization techniques in understanding changes in patterns of dental health behavior between 13 and 17/18 years of age. Methods: Datasets were sourced from the national longitudinal study of young people In Ireland (Growing Up in Ireland). Dental health related variables such as frequency of visits, toothbrushing, dental extractions and fillings and dietary data were selected from the childrens’ cohort at 13 (wave 2, n=7,525) and 17/18 (wave 3, n=6,216) years of age. The data were imported from SPSS Into R in a cleaned data.frame format. R packages ggplot2, reshape2 and dplyr were used to transform and visualize unweighted data. Results: Summary bar plots suggested that the overall pattern of dental health behaviours was similar between 13 years and 17/18 years of age at a population level. There were differences between waves in the overall number of dental extractions (38% at 13 years to 48% at 17/18 years) and dental fillings (17% at 13 years to 21% at 17 years.) However, Alluvial flow diagrams indicated that changes at subject level were not reflected accurately in the overall summary barplots. Similar observations were noted on visualization of dietary variables. Conclusions: Good Data Visualization Techniques Are An Important part of Exploratory Data analysis. Different Visualisations revealed differences in patterns of dental health behaviours at the subject level not immediately obvious at the total population level. These techniques are important when trying to explore, understand and explain cohort data.