Exploding biplots with density axes in Plotly

Sandilands, Delia (2020-12)

Thesis (MCom)--Stellenbosch University, 2020.

Thesis

ENGLISH SUMMARY : Biplots are visual approximations to multidimensional data sets and are very useful in data driven procedures. However, certain issues occur when dealing with a large number of observations or variables. A proposed solution to the former would be moving the variable axes orthogonally to the edges of the plot with an automatic procedure. This unclutters the center of the plot and makes it easier to interpret specific observations. When dealing with a large number of observations the datapoints become indifferentiable from one another and this could be difficult to interpret. This problem was addressed by creating densities on the variable axes. Along with these proposed enhancements, an interactive element to biplot plotting was introduced via the use of Plotly in R. This allows the user to dissect the biplot even further. A web-based Shiny application was built as supplement to this paper, and can be found at https://carelvdmerwe:shinyapps:io/ExplodingBiplots/. Here the user can interact with the data sets, proposed methodology, and functionalities presented in this research.

AFRIKAANSE OPSOMMING : ’n Bi-stipping konstruksietegniek is baie voordelig wanneer veelvoudige data gevisualiseer moet word. Hierdie tegniek het egter ’n paar tekortkominge wanneer datastelle baie veranderlikes of observasies het. Wanneer die datastel baie veranderlikes het, raak die bi-stipping oor vol en dit raak al hoe moeiliker om te onderskei tussen veranderlikes se onderskeie asse. ’n Outomatiese algoritme om die asse te skuif word voorgestel om hierdie probleem aan te spreek. In de geval van baie observasies word dit moeilik om die verskillende gevalle of die waarskynlikheidsdigtheid van die observasies van die verskeie veranderlikes te sien. ’n Moontlike oplossing hiervoor wat insluit die konstruksie van waarskynlikheidsdigthede op elke veranderlike se as is voorgestel. ’n Web-gebaseerde Shiny toepassing was gebou aanvullend tot hierdie artikel, en kan gevind word op die webtuiste https://carelvdmerwe:shinyapps:io/ExplodingBiplots/. Hier kan die gebruiker self die toepassing van die prosesse voorgestel in hierdie navorsing sien.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/109127
This item appears in the following collections: