OpenStreetMap
Great Briain completely mapped in less than a year?
Posted by mannequinZOD on 6 April 2011 in English.Recent ITO world data shows that the current mapping speed is 665 roads/day fixed or added. Assuming that the current mapping speed is kept up, this means the complete road network will be finished in 313 days. Less than a year to go.
http://wiki.openstreetmap.org/wiki/File:GB_stats.png
(100% complete == all roads from the British government Ordnance Service are present and correctly named)
Discussion
Comment from EdLoach on 6 April 2011 at 13:03
Interesting graph (though axis labels are swapped). I'm also assuming that the f(x) has some reference date for where x=0?
I'd also query the "correctly named" claim. With OS Streetview and OS Locator themselves sometimes differing on the name of any particular road, and sometimes both agreeing but not agreeing with what is on the ground, the best that can happen is "probably correctly named, but really could do with survey to confirm".
Comment from chriscf on 6 April 2011 at 17:13
As with any large project, time to completion tends to a constant which is not always zero. We may well end up reaching the point where we are permanently 42 days to completion. Always useful to have a salt cellar handy when daling with these figures.
Comment from mannequinZOD on 6 April 2011 at 18:08
Cheers for your comments. Yes, always take it with a grain of salt. I am sure there are mistakes in OS Locator, but I am also sure there are not many (maybe <0.1%?).
I don't think that we will end up with a permanent 42 days to completion, since a lot of people take the district data from ITO world as an encouragement to improve their area and push it up the rankings. We will end up a bit short of zero, because of the aforementioned errors in OS locator. In the end there is a finite number of streets in GB.
On the other hand, I agree that the entire map will never be completed (POIs, tracks, landuse...), since people are making up new features that can be referenced.
I think the reference date it probably the 1/1/1970, commonly seen as zero in UNIX time, but it doesn't really matter for the calculation.
Sorry about the axis, I'll fix it.
Comment from Vclaw on 7 April 2011 at 17:53
How many of these streets have actually been checked on the ground? Or is it just more people tracing places they have never visited?
The OS data has plenty of errors in it.
Comment from LivingWithDragons on 18 April 2011 at 12:19
There are some districts that are low OSM coverage in the comparison. Please don't just go and map these though, I'm planning mapping events in some of them (probably late summer). My aim is to map them with ground survey AND stir up some new users locally. That will encourage mapping of other features (e.g. footpaths POIs, and house numbers) and to continue updating the map as the world changes next door to them.
Comment from mannequinZOD on 21 April 2011 at 14:19
@Vclaw
I don't know. I have the feeling that some of the areas are rather just traced, seeing the sudden progress of some districts. It's true that OS data has errors, but as long as people do not start to rename streets according to the OS data it is fine. Better to have a misspelled street names than none. Second and third node edits by locals will clear eventual mistakes.
@LivingWithDragons
I agree that it would be better to gather the local community to make it more sustainable for the future. On the other hand, it won't keep people from tracing satellite imagery and fill in names from the OS Data. This has lead to lot's of improvements and can lead to great coverage (e.g. Haiti). Looking at the Gb district data, some are shockingly sparsely mapped.Places like Hastings, City of Plymouth and so on are just at the 50% mark, which is very bad for a populated place.