Longitudinal Data Visualisation
Longitudinal datasets contain a high number of dimensions that require advanced techniques to visually analyse in their entirety. To better appreciate how the number of dimensions, consider a typical longitudinal survey, with thousands of respondents answering hundreds of questions over several waves.
Basic visualisation techniques such as scatter plots are limited to, at most, tens of dimensions, not thousands. A full visual presentation of longitudinal data is a crucial step towards attaining an advanced understanding of the pattenrs within the data and relationships between variables, and complements existing statistical methods of longitudinal data analysis.
To address this problem, ADA have developed a pilot longitudinal visualisation tool called Panemalia. This tool provides a comprehensive visualisation of a longitudinal data set (or cross-sectional, or standalone survey) using a technique known as Parallel Coordinate Plots. This technique is a powerful method of mapping data with a very high number of dimensions down to a two dimensional visual space.
As a pilot study, the tool has then been applied to the first three waves of the Negotiating the Life Course (NLC) longitudinal survey, looking at the labour force participation and relationship status of the survey respondents. To access Panemalia for NLC, you need to be an approved user of the Negotiating the Life Course data (click here to request access).
ADA will continue to add new datasets to the Panemalia system in future. If you would like to include your data in the Panemalia system, please contact ADA at firstname.lastname@example.org.
The ADA GISViz tool has been developed by Kevin Pulo from the ANUSF VizLab at the Australian National University.
- Tools and Technologies for the Social Sciences and Humanities (eResearch 2009)
- Direct Visualization of Longitudinal Data (IASSIST 2010)
We currently recommend the Panemalia tool for use in the early phases of your research, for tasks such as initial exploration and familiarisation, discovery of major trends, and quality control issues such as data cleaning, verification, outlier identification, etc. ADA are interested in exploring additional uses, such as for the verification of analytical results (such as demonstrating the relationship between regression results and the original data), or visualisations for publications. Please contact ADA (email@example.com) if you would like to discuss extensions to the Panemalia system.