I recently found gpt-pilot and wanted a semi-complex project to give it a try with. I decided to port an app in development for MicrobiomeDB to an R Shiny app. That was super fun, and I have no regrets, but it didn't last more than a few days. I think gpt-pilot managed to add a single file input, the correlation coefficient and p-value histograms and filter inputs... that was about it.
An R Shiny app that uploads a file and generates a bunch of pairwise correlations with p-values. If an optional second file is provided it will find the pairwise correlations only between the data in the two files. It will produce a histogram of correlation coefficients, and a histogram of p-values. There are inputs to filter the edges of the network by correlation coefficient and p-value. Currently, the edges are visible in a network and a table. If you provided a single input file you will get a network diagram in the regular way. If you provide two input files you will get a bipartite network with two columns of rodes, where each column represents a file and each node a variable in the file.
Run what I currently have, after installing the package in R, like this: corGraph::corGraph()
Once the app has started, use the provided test files in inst/extdata. If you would like to test the unipartite network, I recommend the file mpg-all-continuous.tsv. If you would like to test the bipartite network, I would recommend the files mpg-cty-hwy.tsv and mpg-year-cyl.tsv.
