Andy has been busy in B72, the heart of Sutton Coldfield, so exactly what does this mean for our map?
Nearly 3500 residential properties,
Almost 300 named retail units,
Nearly 100 commercial buildings,
All drawn in and with full address details including the full postcode!
Its taken a little over two months, a half hour here and there out walking and cycling the streets to collect all the address data; house numbers, retail names etc. Many more hours at the computer drawing each of nearly 4000 seperate buildings and in the case of residential properties their associated gardens.
This means that B72 is essentially fully mapped out. Nearly 50,000 objects in all (nodes, ways and relations). Of course we could find more to map even now but the current result is as detailed and rich a dataset/map as almost anyone should ever require.
You can browse the map here or search out an address in B72 using nominatim.
In undertaking an exercise such as this there are always lessons to be learnt. Here’s just a few that may help and spur on others to map their areas in full detail.
- Tools that orthogonalise buildings (hit “Q” in JOSM or Potlatch2) are a godsend, also the duplicate and node merge (“M”) facilities when it comes to creating multiple buildings of the same shape and size.
- Most traditional retail units have a front facing length of 20 feet (6.1m). Very handy when dividing up long runs of shops.
- Postcodes were added using the OS OpenData CodePoint dataset. Postcodes are unique to a particular street (or in some cases particular address) with postcode changes along a street normally occurring at a connecting side street or other natural break (public footpath perhaps). In a handful of cases its impossible to know if a building opposite a side street is in one postcode or another and in these few instances the postcode has been left as a “?” until it can be verified.
- When looking at all the property boundaries you find a number of other features that are easily overlooked. Local electric substations being a good example.
- There is no easy way to deal with properties with vertically mixed use. For example flats over shops. A longer term solution is needed for these.
- Property addresses in private gated roads are difficult to obtain unless you are granted access.
- To show all business names in the retail heart of a town or city you really need an extra zoom level (z19) or a method of better dealing with icon and name conflicts on the rendered map.
So what’s next, well of course B72 is just one small area containing just 242 unique postcodes. Birmingham postal region alone has nearly 42,000 unique postcodes so there is a huge amount to do even for our local mapping group and without more help it will take many years. Please help by joining the project, every property and address you add to the OpenStreetMap project means more information available to West Midlands individuals, organisations and businesses.
More: Mappa-Mercia OpenStreetMap user group
Contact us: info (at) mappa-mercia.org
B27- how’s that for palindrom-ness- shouldn’t be too far behind (like a few weeks)- but I don’t think I’ll ever have the patience to enter all the garden boundaries
Excellent work, and very inspiring – when the weather gets a bit nicer I might attempt to do a little bit of this close to my home (I don’t think I’ll manage the whole of a postcode area, or even sector, but possibly I’ll do the whole of my postcode!)
Fine work, sir!
I’m doing a similar job in Cambourne, Cambridgeshire:
I’d agree with all your points above too, based on my experience.
In Cambourne, I’ve added driveways to garages and footpaths to front doors too, as it became obvious early on that without these, it often wasn’t at all clear how to access a particular property, which I think is one of the main uses of this kind of mapping. (See, for instance, nos. 5 and 7 on Alder Drive in the above linked map. Without indicating fences, driveways and footpaths, they would just be sitting marooned in between several surrounding roads).
This appears to be less of a problem in B72, as the roads are much straighter and more widely spaced, so it’s pretty obvious which houses are accessed from which streets.
It’s just a shame that the village was still under construction when the Bing imagery was made, so we’ll have to wait for complete mapping of the last parts to be built.
How can you do that without using copyrighted data from other mapping? Even if you walked down the street writing down the house numbers you’re not going to know the shapes of the houses without either OS mapping or satellite data or both.
Something like this would be wonderful for delivery drivers, especially in the countryside. But I thought the obstacle to producing such a thing was that you’d need the PAF, the OS maps, Google, etc etc to accomplish it.
This is done by editing the crowd-sourced map “OpenStreetMap” (OSM). Microsoft has allowed OSM contributors to use their bing aerial imagery and the centre points of postcodes is now available as opendata (from both the OS and the Office of National Statistics).