It’s no seekrit that I :heart: Hilbert curve heatmaps of IPv4 space. Real-world IPv4 maps (i.e. the ones that drop dots on the Earth) have little utility, but with Hilbert curves maps of IPv4 space many different topologies can be superimposed (from ASNs to—if need be—geographic locations). Plus, there’s more opportunity to find patterns by keeping CIDRs naturally close to each other.
The Measurement Factory created the [`ipv4-heatmap`](http://maps.measurement-factory.com/) command-line utility back in 2007 and there have been some tweaks and expansions to it by others over time. I wanted to use these IPv4 heatmaps in the [National Exposure](https://community.rapid7.com/community/infosec/blog/2016/06/07/rapid7-releases-new-research) report I worked on with @todb & @jhartftw at @Rapid7 but _cannot stand_ the built-in red-blue color scheme, especially when there’s [viridis](https://cran.r-project.org/web/packages/viridis/vignettes/intro-to-viridis.html) available. So, I [forked the code](https://github.com/hrbrmstr/ipv4-heatmap) and added both viridis and [colorbrewer](colorbrewer2.org) palettes to it as command-line options.
Here are two examples (the results of the National Exposure study), one using viridis and one using the colorbrewer `rdbu` palette:
You specify the palette with `-P palette` and can invert the order of any palette with `-i` and the chosen palette will also be used in any legend you add the visualization.
Since these 4096×4096 files are a bit big, you can hit up [this dropbox link](https://www.dropbox.com/sh/wqyly8ewxeko5jn/AAC5bHIpQTuxWGBPYzMqceLQa?dl=0) to see a “gallery” of the various forward and reverse palettes.
The palette selection code is a bit brute-force at the moment, mostly due to the fact that I’m planning on a C++ port of the code and eventual inclusion of the Hilbert heatmap functionality in the [`iptools`](http://github.com/hrbrmstr/iptools) package.