[Talk-GB] Bristol - a quick (and surprising?) statistic...

[Tim Francois]

If you do find mistakes in the OSL data (and any other OS OpenData
products), please do let OS know - email
customerservices at ordnancesurvey.co.uk. In addition, add the entry to the
Catalog of Errors: http://wiki.openstreetmap.org/wiki/Catalog_of_Errors.

As for your point: "Tim's analysis is very useful in pointing out areas for
attention, but the reality is not as bad as the figures suggest."

As others have pointed out, that's the reality in your area of 46 roads -
there are 9282 roads covered by my Bristol analysis: the reality in any area
will differ greatly - just see the ITO table for some surprises!!. I have
found, though, that many, many roads here are named just as they are on the
street signs: so shortenings such as rd, pl, ave etc seem to be common
place. This is logged as an 'unfound' road by the script as it obviously
doesn't match the OSL data. So far, I've re-surveyed around 50 roads or so,
and found all but 2 to be errors in OSM - I've logged the errors in the
Catalog if anyone's interested.

Thanks
Tim

On Mon, Aug 2, 2010 at 10:15 PM, Richard Moss <richard at richardmoss.co.uk>wrote:

>
>
>
>  On 23 July 2010, Tim François wrote:
> ...
> More info here:
> http://wiki.openstreetmap.org/wiki/User:Tiiiim#Bristol_-_Missing_Roads
> Perhaps the figure won't be surprising to many, but I thought it was pretty
> large! Is this representative of other similar sized cities/towns? (I know
> OSL isn't 'perfect', but its better than nothing....)
>
> ...
>
> I was shocked - here in Cambridgeshire, we thought we were getting close to
> 100% , particularly in the City and South Cambs, so I've had a look at my
> own village to try and understand what's going on.
>
> On the Royal Mail list of postcodes for the village, there are 46 streets,
> and in OSL, 10 of them are wrong - i.e. 78% correct, which is a bit worse
> than the 82.98% figure for South Cambridgeshire given at
> http://www.itoworld.com/product/data/osm_analysis/
>
> This is a list of the 10 that are wrong:
>
> Name in OSL             Name on OSM                     Comment
>
> Chapman's Way   Chapmans Way            2 road name signs, both say
> Chapmans: OSM correct
> Hill Road               Mill Road                       Mill Road is
> correct
> Unwin's Lane    Unwins Lane                     2 road name signs say
> Unwins and 1 says Unwin's
> High Street             High Street (Over)      Useful distinction from
> nearby High Streets in
>                                                        Swavesey and
> Willingham, especially when these names
>                                                        appear in the like
> of www.cyclestreets.net . I'm not
>                                                        going to change what
> has been put on OSM
> Station Road    Station Road (Over)     ditto
> The Old Wood Yard       The Old Timber Yard     Whoops - my mistake.
> Careless adding of a new development.
>                                                        Sign says The Old
> Woodyard (all one word) so both wrong!
>                                                        I will correct
> Whines Lane             Whine's Lane            There are 2 road name
> signs. One of each
> The Doles               The Doles (East)                ) OSM reflects
> signs on the ground. OSL agrees with the
> The Doles               The Doles (West)                ) Royal Mail street
> list.
> Randalls Lane   Randall's Lane          Street sign says Randall's, as OSM
>
> (apologies if the formatting comes out wrong)
>
> In reality, there's only one error in OSM, and 45/46 is 97.8%. I don't want
> to get into a debate about apostrophes!
>
> You could then note that there are various farm tracks outside the village
> which have names in OSL, missing on OSM and not in the Royal Mail postcode
> list: Long Lane, Ouse Fen Road or Bank, Long Holme Drove, Long Drove. There
> is no evidence for these names on the ground. Some have been added to OSM
> from local knowledge: Chain Road, Gravel Bridge Road. Then we are perhaps at
> 47/52 = 90.4%
>
> So I am heartened.  Tim's analysis is very useful in pointing out areas for
> attention, but the reality is not as bad as the figures suggest.
>
> Richard
>
>
>
>
>
> _______________________________________________
> Talk-GB mailing list
> Talk-GB at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-gb
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-gb/attachments/20100804/0a7a29ff/attachment.html>