This Shiny application aims to provide an interactive experience to the awesome corrplot R package. Correlation is one of the most commonly used metrics to depict statistical relationships between two variables. It's often used in exploratory analysis, feature selection and many other modelling steps.

Currently we provide a list of built-in datasets sourced from various R packages. You can also upload your own in any delimited flat file format.

You can use checkboxes to pick only the variables you want to see. Furthremore, you can drag variables around to manually reorder the columns. This will have an obvious impact if you choose to see correlation in their original order. However it can also affect some reordering algorithms such as AOE.

The reordering by checkbox updates the plot in realtime while the selectize style only udpates when you drop the selection. When the server is powerful enough, the realtime update can be very helpful. Otherwise, it will seem visually disruptive.

The cor function in stats package is used thus you have the freedom to choose from pearson, kendall, and spearman correlation and some control over how to treat NA values.

We have include abilities to generate most examples in corrplot vignette. In particular, you can make statstical inference on correlation significance and/or confidence intervals.

Thanks goes to corrplot and sortList example for inspiration, and Developing Data Products on Coursera for motivation and this great opportunity.

MIT