PCA Explorer

All multivariate data, irrespective of their origin (experiment & field), share common representation. In the majority of cases, such data (often belonging to multiple ‘classes’ or ‘groups’) are represented in the form of a 2D matrix, with features (or dimensions) on one side and corresponding data points on the other. Due to the limitation of representing multiple dimensions on a 2D plane, informative visualization of the same becomes a challenge. Consequently, the most logical approach to visualize such multidimensional data is using techniques like Principal Component Analysis, which decomposes the multiple features into two most discriminating dimensions, followed by plotting the same on an X–Y plane scatter. PCA Explorer is a free web based tool for generating Principal Component Analysis (PCA) plots, interactively query them and apply statistical tests between the available classes. The input data can be uploaded as a simple tab delimited abundance matrix along with multi-class metadata information (in a separate file). The tool has inbuilt modules to filter and normalize the input data and simultaneously view the output plots.

Link : https://web.rniapps.net/webigloo/pca/