[Talk-GB] Summer quarterly project

Dan S danstowell+osm at gmail.com
Tue Sep 13 22:36:50 UTC 2016


Town-centre blitzes can lead to out-of-date data very quickly. Town
centre data is better to have in OSM if it has maintainer(s). Jerry
acknowledged this, but I still would like to register a concern about
that!

Best
Dan

2016-09-13 18:31 GMT+02:00 Paul Berry <pmberry2007 at gmail.com>:
> +1 for the town centre blitzing. Even well-mapped city centres change all
> the time (shops opening and closing for one) and there are plenty of
> examples of places you'd think would have more detail than they actually do
> (eg Leeds is still pretty poorly mapped for POI and businesses, despite my
> best efforts). This is of course largely a reflection on how strong the
> local mapping community is. London, Cambridge and Nottingham spring to mind
> as exceptionally detailed urban areas.
>
> I think it'd have a bigger impact on the OSM people actually use. But, even
> if we don't go for this for the next Quarterly, it's always there as a
> "background" mapping task for all.
>
> Regards,
> Paul
>
> On 13 September 2016 at 15:46, Brian Prangle <bprangle at gmail.com> wrote:
>>
>> The idea of adding opening hours and lots of fhrs data to existing OSM
>> data is not one I personally find attractive.  But I do like Jerry's idea of
>> blitzing town centres that are poorly mapped. Our own experience in
>> mappamercia recently in Kidderminster where we had  a summer Saturday
>> mapping meetup shows that a handful of mappers  with photo surveys can
>> completely transform the map. Add that to frhs data and we can have an even
>> better map with address data. It will also get us out to new areas for
>> mapping and increase the opportunity for community-building amongst
>> ourselves eg "I'm going to xyz this Sat - anyone care to join me";
>> contacting local mappers to see if they want to join in; contacting the
>> local chamber of commerce etc.
>>
>> I also like the idea of improving the road network generally - road
>> alignment(very poor in some areas) speed limits, lane counts,turn and
>> destination lanes information etc. - gets us back to the basics and possibly
>> impacts more data users
>>
>> On balance I prefer the town centre blitz approach as it offers more scope
>> for community building
>>
>> regards
>>
>> Brian
>>
>> On 13 September 2016 at 12:32, SK53 <sk53.osm at gmail.com> wrote:
>>>
>>> My comments on both suggestions:
>>>
>>> Speed Limits: a little bit boring, BUT there are some relatively
>>> achievable targets. For instance getting all primary & trunk roads with
>>> speed limits. There are areas of the country where none of these roads have
>>> limits, but even in well mapped places there is a considerable amount of
>>> simple tidying up (missing speed limits on roundabouts or short sections)
>>> which can be done. A further advantage is that major roads are also more
>>> likely to have Mapillary/OpenStreetView coverage. Additionally things like
>>> number of lanes, availability of pavements etc can be added as well whilst
>>> reviewing speed limits. I noticed this a few weeks ago because back in
>>> September last year I drove to Bewdley & waypointed changes of speed limits
>>> on the A456 from Hagley to Kidderminster.
>>>
>>> One additional caveat is that speed limits on the narrower roads are
>>> changing a lot: national speed limits to 50, 50 mph down to 40 mph, etc.
>>> I've noticed this particularly along the A606 as travel this by bus about
>>> once a year when I take more mapping notes.
>>>
>>> I have put a map based on this Overpass-turbo query on Flickr for trunk
>>> roads missing speed limits, and one for primary roads in the East Midlands
>>> here. The latter query returns too much data for the whole of the country
>>> but can be tailored by changing the area part of the query.
>>>
>>>
>>> Food Hygiene data. This would be in two forms: enrichment of existing OSM
>>> data (primarily with addresses); and surveying areas which have lots of FHRS
>>> data but little in OSM.
>>>
>>> The former is a valuable, but not particularly gripping activity. IIRC
>>> the FHRS data covers somewhere between 10-15% of total postcodes, and just
>>> having one address in a postcode can help resolve many adjacent ones. Two
>>> addresses and one can infer properties of how addresses are allocated on a
>>> road.
>>>
>>> Last year Peter Reed wrote a long series of blog posts about retail data
>>> and used Super Output Areas to predict volumes of missing data from OSM.
>>> Last year I targeted Melton Mowbray, Coalville, Havant and Chichester for
>>> mapping of the town centres based on this data. More recently I've done
>>> Hoylake & New Brighton. I'd hoped to have a look at Hyde, Tameside at the
>>> weekend, but was too tired by the end of the field meeting. Most towns in
>>> Greater Manchester are ripe for this kind of mapping: Oldham, Rochdale,
>>> Hyde, Denton, Ashton-under-Lyme and many others. In the past I have used a
>>> set of postcode centroids denoting places with missing data to help target
>>> the mapping. More recently I have munged the FHRS data by distributing all
>>> places sharing a postcode on a circle of 10-20 m radius and created GPX
>>> files for particular areas: with each FHRS category having a different
>>> symbol.
>>>
>>> A town the size of Melton Mowbray took around 90 minutes to do a photo
>>> survey. Adding the data to OSM rather longer. Stockport, another, larger,
>>> retail centre, which I have now surveyed in 2015 & 2016, took about 3 hours
>>> altogether. Some of this was duplicated, and in part was because the
>>> Merseyway Shopping Centre closed before I got round it on my first visit. A
>>> second visit is useful because one inevitably notices anomalies which
>>> require investigation when entering the data. My strategy is to take photos
>>> and a limited number of notes or audio files, and therefore maximise mapping
>>> time. This is all based on FHRS having all the other relevant data, which
>>> doesn't work everywhere.
>>>
>>> In summary using FHRS data enables a fairly targeted approach to mapping
>>> town centres. It greatly helps in assembling address data in such places
>>> and, of course, adds to the detail. For the mapper it's quite rewarding
>>> because one can quickly see the impact. Although shops can change a lot,
>>> having the much less ephemeral address data ensures this is not a Red Queen
>>> ('running to stay still') task.
>>>
>>> Jerry
>>>
>>>
>>>
>>>
>>> On 13 September 2016 at 09:25, Ed Loach <edloach at gmail.com> wrote:
>>>>
>>>> Paul commented on John's suggestion:
>>>>
>>>> > Speed limits would be a good one, although impossible to armchair-map
>>>> > unless you know
>>>> > something I don't. Also, would it stem the tide of useless speed limit
>>>> > notes from Navmii GPS users?
>>>>
>>>> I can't guarantee it would stem the tide of Navmii speed limit notes,
>>>> but I added lots of speed limits locally when Skobbler were creating MapDust
>>>> notes and it seemed to stem that tide.
>>>>
>>>> I like Robert's fhrs suggestion - I keep meaning to cross check what is
>>>> and isn't mapped against the fhrs list but haven’t got around to it. Making
>>>> it a project might spur me to do so.
>>>>
>>>> Ed
>>>>
>>>>
>>>> _______________________________________________
>>>> Talk-GB mailing list
>>>> Talk-GB at openstreetmap.org
>>>> https://lists.openstreetmap.org/listinfo/talk-gb
>>>
>>>
>>>
>>> _______________________________________________
>>> Talk-GB mailing list
>>> Talk-GB at openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk-gb
>>>
>>
>>
>> _______________________________________________
>> Talk-GB mailing list
>> Talk-GB at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk-gb
>>
>
>
> _______________________________________________
> Talk-GB mailing list
> Talk-GB at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-gb
>



More information about the Talk-GB mailing list