Visualization of Time Series Data Change by Statistical Shape Analysis

Abstract

Visualization is very effective when we analyze the big data. When we handle with time series data for a long period, especially visualization is needed. In the paper, we shall illustrate change of the whole trend on time series data as shapes. The method we used is statistical shape analysis which can extract the Affine and non-Affine transformations from the deformation. The method is also helpful to see a local movement of each data sample, compared to other neighbors. In the paper, we illustrate the whole trend change using the Indonesia province comparison concerning the total fertility rate and the under five mortality rate. From the visualization, we found that the relationship between the data have been relatively stable and shows a linear relationship, finally in 2012.