OS Locator
OS Locator (OSL) is one of the datasets released as part of the Ordnance Survey Opendata releases in the UK. It is a file with a list of road names and positions for the UK. Although the file does not give road geometry, it does give a rough location for every street. This is good enough to use as check list / comparison dataset, in tools for Quality Assurance of OpenStreetMap and for measuring completeness.
Contents |
Comparison tools
There are several tools for comparing OS Locator streets data with OpenStreetMap.
- OSM and OSL differences tileset - From ITOworld, this is available for use as a background layer with editors. Follow the link for instructions on how to use an editor with these tiles.
- OSM and OSL differences analysis ([1]) - Stats and comparison of number of differences remaining, on a borough-by-borough basis.
- OS Locator Musical Chairs ([2]) - A browseable OpenLayers interface to the results of a smart fuzzy-matching musical chairs algorithm, highlighting disagreements between OS Locator and OSM and Locator street names that are altogether missing from OSM.
- oslVosm - A php script which you can run on your own area.
Interpreting the differences
 |
OS Locator contains errors. You should never just change street names in OpenStreetMap to match OS Locator names. |
You should never just change street names in OpenStreetMap to match OS Locator names. That's really not the idea at all! By doing this, you could be using information with errors and overwriting information which has been properly surveyed by people out on-the-ground. Even if the surveyed OSM data looks like an error, you should assume that data is correct unless you can get there on the ground to check.
It is also quite widely regarded as unhelpful to copy streetnames from OS Locator for places you have never visited. Meanwhile some people are doing this in some areas. If you do this you should definitely use the source tag (and probably on the elements themselves) :
source:name=OS_OpenData_Locator
If this tag is used on elements (as opposed to changesets) then mappers should be bold and remove the source tag or replace it, when they have surveyed and verified the street name.
Discussion: Vclaw's diary, chriscf’s diary,discussion on talk-gb , and no doubt many other places (TODO: link here)
TODO: More general notes on interpreting the differences and corrective actions to take with editors.
Create GPX file of OSL data
Useful for:
- JOSM (can be opened as a new layer)
- Google Earth
- oslVosm (automatic OSL and OSM matching script)
Prerequisites
This method has been tested using Ubuntu 9.04. It requires the following:
- grep (standard in Linux, and also in cygwin)
- gpsbabel (in the standard repositories)
- A spreadsheet program. OpenOffice.org Spreadsheet is fine.
Get OS Locator data
- Go to https://www.ordnancesurvey.co.uk/opendatadownload/products.html and request a download link for the OS Locator product.
- Once the link has been emailed to you, download the zipped file, then extract the file OS_Locator2010_1_Open.txt.
Process OS Locator data for your area
- First, extract from the data file the items which are in our area. In this case, Bath and North East Somerset. Save this to a text file called BANES.txt.
grep "Bath and North East Somerset" OS_Locator2010_1_Open.txt > BANES.txt
- BANES.txt now contains all the streets within the BANES area, one per line with corresponding OSGB Grid References for their locations. Each field within a line is separated by colon.
Convert Data into Useful GPX file
- Open the file BANES.txt using Excel or OpenOffice Spreadsheet, making sure to choose the ":" as the field delimiter.
- In cell Q1, add the following:
=CONCATENATE(LEFT(O1,2), " ", MOD(C1,100000)," ",MOD(D1,100000))
- Copy this down for the rest of the rows. This provides a 10 figure grid reference, which will be fed to gpsbabel and converted into lat/lon pairs. To do this, first save or export the spreadsheet as a tab-delimited file, BANES.tab.
- Next, copy and past the following gpsbabel stylesheet into your favourite text editor, and save as OSGB.style.
# Format: Import (Excel-modified) OS Locator street name centres # Author: D Dixon # Date: 8 April 2010 # DESCRIPTION Import (Excel-modified) OS Locator street name centres EXTENSION txt # # FILE LAYOUT DEFINITIIONS: # FIELD_DELIMITER TAB RECORD_DELIMITER NEWLINE BADCHARS DATUM OSGB36 # # INDIVIDUAL DATA FIELDS, IN ORDER OF APPEARANCE: IFIELD DESCRIPTION, "", "%s" # Street name IFIELD SHORTNAME, "", "%s" # Street ref IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD IGNORE,"","%s" IFIELD MAP_EN_BNG,"","%s %5d %5d" # OS Grid reference
- Now, run the following command to process BANES.tab and convert it into a GPX file.
gpsbabel -w -i xcsv,style="OSGB.style" -f "BANES.tab" -o gpx -F "BANES.gpx"
Cut down data to Bath City area only
Currently, BANES.gpx has all the roads in the Bath and North East Somerset area. I only care about the roads actually within Bath, so we need to cut down the GPX file and extract only those roads which are within the city. This is done using the polygon filter in gpsbabel.
- Using http://www.openstreetmap.org/, determine the appropriate bounding box for your desired area.
- Using this data, create a file called BATH.bb, and insert the bounding box in the following way:
51.4249453 -2.4529905999999997 51.337054699999996 -2.4529905999999997 51.337054699999996 -2.2772094 51.4249453 -2.2772094 51.4249453 -2.4529905999999997
- Those are the corner positions in lat/lon for the bounding box of Bath. Notice that the last line is the same as the first - this is required to close the box. Now, run the following to extract only the roads within Bath from BANES.gpx.
gpsbabel -w -i gpx -f BANES.gpx -x polygon,file=BATH.bb -o gpx -F Bath_Streets.gpx
You now have a GPX file which can be loaded into JOSM or Google Earth.
