Package to build heatmaps for economic analysis, v. 0.4.
heatmapEco is a R package that handles both the creation of heatmaps - colour-coded, high-dimensional data visualizations - as well as any data munging needed to produce the visualization.
This package is an alpha build. Bugs are possible, and major changes to the interface should be expected in the future.
Why heatmapEco over other heatmap-making packages? This package recognizes that heatmaps can also be used to plot the relationshio between a categorical/factor variable and a continuous variable, adequately discretized. The goal is to create a package with an easy-to-use interface that processes all the discretization, aggregation and relabelling involved with a heatmap automatically.
heatmapEco also has a companion interface in Stata, just called heatmap. The interface communicates with the R package for a more robust program than Stata's own heatmap program.
1) Of the R package
Pull the repo to get the source package, heatmapEco_xxx.tar.gz. Install it from source in R, or run heatmapSetup.R from the repo to automate installation. To run heatmapSetup.R, make sure to input the directory to which the repo was downloaded as the command line argument.
Upload to CRAN is coming eventually.
2) Of the Stata package
Run sysdir in Stata. Copy the ADO and STHLP files in subdirectory stataHeatmap to one of the folders listed by that command (the "PERSONAL" or "PLUS" folder is advised). Reboot Stata; the command should be loaded afterwards.
The Stata help file goes through more details about the installation of the R package if you are an R novice.
Upload to SSC is tentative.
Check out the slides in docs/heatmaps.pdf for an example of what can be done with heatmaps.
For more information read Tom Cui and Eric Zwick, [NBER ARTICLE HERE]
The subdirectory /RHeatmap has the project file for the R package, so with RStudio you can start editing right away.
There is only one ado file for Stata Heatmap, which can be edited directly.
v 1.0 should include a much more "functional" interface in both the Stata and R interfaces, where settings to an axis are called as suboptions to a main option for each axis.
Allow easy aggregation of data on just one axis, not both axes, and display those aggregated data as sideplots to the heatmap.
Tom Cui: E-mail Tom.Cui [at] chicagobooth.edu.
heatmapEco is licensed under GPLv3.