About

I am a physics student from Karlsruhe. Currently writing my bachelor thesis, which is not a good time to get interested in a new hobby. Except nerdy stuff such as science and computing I also like doing sports. I do Judo, go regularly jogging and I like cycling, so I will try to reach most of my geohashes by bycicles.

My geohashing implementation in emacs

https://github.com/elimik31/geohashing_el

Analyses

Using python with the modules numpy and matplotlib I ran some monte carlo analyses with random numbers to determine the probability to find geohashes at different distances. The first plot show the probabilty to find a geohash in a certain distance intervall (between to radiuses) and the second histogram is an integrated version of the first one, showing the probability to find a geohash with a distance smaller than a certain radius. The errorbars in the first histogram are calculated from the standard deviation of the entries in each bin.

First I used only the graticule of my coordinates [49, 8], but then I started also started to consider the graticule [48, 8], as Karlsruhe lies on the border to this graticule. Later I also added to further graticules, [49, 7] and [48, 7]. As I was interested, how this would affect my histograms, I put all of them on one figure.

The curves in the left plot can be understood when you understand that the probabilty is proportional to the circumfance of the distance cirle. As long as all of the circle fits into the first graticule, the probability rises linearly with the radius. Then, it begins to fall, as more and more of the cirle starts to get cut off at the sides and so on...

File:Graticule numb dependence.svg