Visualizing OpenStreetMap history – part 2
Thanks to the hard work of Vincent we now have a database with the average version and average age of nodes per grid cell. Now we can start to get a feeling of the data. I’ve rendered both datasets for Amsterdam on a 10x10m grid. In general the idea is that red is bad and green is good. The younger a node is, the more likely it reflects the current situation in the real world. Also the higher the version number, the more people have been looking at that node and corrected it.
At least that is the theory Martijn tries to work with.
The age in days for OpenStreetMap nodes
You can see a clear pattern emerging here, apparently there are areas which get mapped in a short period of time and never touched again. The white squares just didn’t have any nodes in it, so they contain no data.
The version of OSM nodes
It is more difficult to see patterns here. Whereas with age in days the number gives you an instant grasp of the meaning, the version number is less obvious. We probably need to do a statistical analysis of version numbers in OSM. This way we can attach some (relative) meaning to the average version number.