This occurrence of the bi-annual corruption of the space-time continuum (i.e. changing to/from standard/daylight time) in the U.S. caused me to make a slight change to the code [from an older post](https://rud.is/b/2014/09/23/seeing-the-daylight-with-r/). The `daylight()` function now auto-discovers the date and location information (via [telize](http://www.telize.com/)) from the caller, which means all you have to do to get a plot like this:
is to source the [new gist](https://gist.github.com/hrbrmstr/e435d4fa0c31b8e1a9d0) like this:
devtools::source_gist(“e435d4fa0c31b8e1a9d0″, sha1=”64e859227266dc5f9008b3b3959a19fea373fee6”)
Remember that you should verify any code before blindly `source`ing it (in R or anywhere else) and make sure to use the SHA1 hash so you know you’re sourcing the proper code (and not potentially being pwnd).
Note that the granularity/accuracy of the geolocation is only as good as the Telize service (which uses MaxMind). The fact that this shows Vermont instead of Maine should make you all think thrice about trusting IP geolocation in general, especially you world-mapping cybersecurity folks.
Sadly, the darkest of days is still yet to come.
One Comment
I like this plot very much. Unfortunately I still have to use version 1, because the telize server seems not to be so powerful these days. Therefore I always get an timeout error.