place=? An oldie but no past conclusion.

classic Classic list List threaded Threaded
35 messages Options
12
Reply | Threaded
Open this post in threaded view
|

place=? An oldie but no past conclusion.

Warin

Hi,

I came across Quobba Station and Canarvon where Canarvon is a 'village' and Quobba Station is a 'town'. I know Canarvon is larger than Quobba Station!
So I re-tagged Quobba Station as 'village'. 
But I wondered on it.. so looked up the OSMwiki .. not much help... the Australian tagging guidelines ... errr not really. 
I then considered getting all the Australian OSM data on places with the population data, 
Got the cities data fine, but the towns data is too large a single bite and the server objected. Fine, I worked on the city data. 
Some 90 are set as cities... I'll ignore those above 10,000 people and list the others here so you have an idea of those that maybe reclassified as 
'towns' under my proposal. If a place is close to the 10,000 mark and there are no others around that location then I'd consider it a city, but other wise a town. 
Charters Towers    8,234
Charleville        4,700
Caloundra        3,550
Winton        1,337 
I know Winton ... it is smaller than Longreach (both in population, about 3 times, and number of pubs).. yet Longreach is not tagged a city?
 
Conclusion: there is a significant error in the relative ratings between places - even ones that are not that far apart! 

The situation with towns and villages is more numerous! 
The server objected to my bulk download ... so I'll do that in bits later ... unless there is no point - that is if there are strong objections here?
Little point in doing the large bit of work if there will be no outcome.

So below is a small attempt to clarify and simplify the situation in Australia.
From the OSM wiki I get the following use of occupied places
By population.
city>100,000>town>10,000>village>200>hamlet>100

The present Australian use appears to be roughly 
By population.
city>10,000>town>1,000>village>100>hamlet>10

I think that is reasonable. 
The difference between the two is that Australia has a smaller than 'average' population density, 
so smaller places have more facilities due to the distance involved to get to the nearest larger place. 
For example - Australia is about the same size and mainland USA .. but 1/10 th the population.. 
so it stands to reason that the Australian population density would be about 1/10th .. so a 'town' would be about 1/10th too.
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.

Problems... ?
 In large centres like Sydney and Melbourne some parts would be judged as 'cities' in their own right ... 
not certain if that is a problem or not? Comments? I am more concerned with the country side, rather than the messy cities. 


Are there any objections/comment/other ideas to the above ?
---------------------
I have read the past posts on this ...
https://lists.openstreetmap.org/pipermail/talk-au/2008-December/001079.html
but I could draw no reasonable conclusion. 
There was a suggestion that the number of pubs be used ... which I think is quite Australian, 
I use it to judge safety when parked .. less than 3 pubs = safe.. everyone knows everyone. More than 2 - cover and lock up.
By pubs
city>20>town>3>village>1>hamlet 

_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? An oldie but no past conclusion.

Christopher Barham

On 03 May 2016, at 14:22, Warin <[hidden email]> wrote:


<SNIP>
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.
</SNIP>

City is not just a function of population - It’s can also be a political appointment/status? - e.g. Charters Towers and Redcliffe are cities : https://en.wikipedia.org/wiki/List_of_cities_in_Australia


C.

_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? An oldie but no past conclusion.

cleary
In reply to this post by Warin
 
From what you have written, I think you are probably right to change classificaiton of Quobba Station to village rather than town.
 
In regard to smaller rural centres, I have also been wondering about classifications. In South Australia, some rural properties constitute their own official location with their own allocated postcode. In one case the farm is known as "Oakvale" but the official Government-approved name for the location with defined boundaries and its own allocated postcode is "Oakvale Station". I had originally labelled it as place=farm but the OSM WIki says that if the farm is not part of a larger settlement, then it should be labelled as "isolated_dwelling". Since it is its own suburb and not part of a larger settlement, I have labelled it as place=isolated_dwelling.
 
"Hamlet" can be a small population but does not necessarily have a central location- thus Mossgiel in NSW has a defined area and postcode with scattered farms in the area but there is no centre with shops or services - thus it is tagged as place=hamlet. Once a place has shops, pub, other services etc. I think it qualifies as a "village".  But two places could be villages even though one is a fair bit larger than the other - if you can get food and fuel etc. but it's not big enough to be a town, then "village" is a good classification.
 
In NSW, "city" is an official Local Government Area status approved by Government so that some larger towns (and their surrounding areas) are designated as cities e.g Albury, Waga Wagga, Griffith and locations within their city boundaries are officially "suburbs". The "suburb" with the same name as the LGA is usually also identified as "city". In Sydney, some major suburban centres such as Parramatta are identified by Government as cities. I think the official classification works fairly well in NSW  and is useful guide for tagging in OSM but not sure about other states.
 
Then there are areas that have defined boundaries as "suburb" and designated postcodes but have no permanent population as far as I know (such as Jagungal Wilderness in NSW or Danggali in SA). Not sure how to classify them. "Locality" is used for much smaller places that have no boundaries, postcode or population and seems not the right tag for larger areas with defined boundaries.
 
 
 
On Tue, May 3, 2016, at 10:22 PM, Warin wrote:


Hi,

I came across Quobba Station and Canarvon where Canarvon is a 'village' and Quobba Station is a 'town'. I know Canarvon is larger than Quobba Station!
So I re-tagged Quobba Station as 'village'. 
But I wondered on it.. so looked up the OSMwiki .. not much help... the Australian tagging guidelines ... errr not really. 
I then considered getting all the Australian OSM data on places with the population data, 
Got the cities data fine, but the towns data is too large a single bite and the server objected. Fine, I worked on the city data. 
Some 90 are set as cities... I'll ignore those above 10,000 people and list the others here so you have an idea of those that maybe reclassified as 
'towns' under my proposal. If a place is close to the 10,000 mark and there are no others around that location then I'd consider it a city, but other wise a town. 
Charters Towers    8,234
Charleville        4,700
Caloundra        3,550
Winton        1,337 
I know Winton ... it is smaller than Longreach (both in population, about 3 times, and number of pubs).. yet Longreach is not tagged a city?
 
Conclusion: there is a significant error in the relative ratings between places - even ones that are not that far apart! 

The situation with towns and villages is more numerous! 
The server objected to my bulk download ... so I'll do that in bits later ... unless there is no point - that is if there are strong objections here?
Little point in doing the large bit of work if there will be no outcome.

So below is a small attempt to clarify and simplify the situation in Australia.
From the OSM wiki I get the following use of occupied places
By population.
city>100,000>town>10,000>village>200>hamlet>100

The present Australian use appears to be roughly 
By population.
city>10,000>town>1,000>village>100>hamlet>10


I think that is reasonable. 
The difference between the two is that Australia has a smaller than 'average' population density, 
so smaller places have more facilities due to the distance involved to get to the nearest larger place. 
For example - Australia is about the same size and mainland USA .. but 1/10 th the population.. 
so it stands to reason that the Australian population density would be about 1/10th .. so a 'town' would be about 1/10th too.
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.


Problems... ?
 In large centres like Sydney and Melbourne some parts would be judged as 'cities' in their own right ... 
not certain if that is a problem or not? Comments? I am more concerned with the country side, rather than the messy cities. 



Are there any objections/comment/other ideas to the above ?
---------------------
I have read the past posts on this ...
https://lists.openstreetmap.org/pipermail/talk-au/2008-December/001079.html
but I could draw no reasonable conclusion. 
There was a suggestion that the number of pubs be used ... which I think is quite Australian, 
I use it to judge safety when parked .. less than 3 pubs = safe.. everyone knows everyone. More than 2 - cover and lock up.
By pubs
city>20>town>3>village>1>hamlet 
_______________________________________________
Talk-au mailing list
 

_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? An oldie but no past conclusion.

Warin
In reply to this post by Christopher Barham
On 4/05/2016 12:50 AM, Christopher Barham wrote:

On 03 May 2016, at 14:22, Warin <[hidden email]> wrote:


<SNIP>
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.
</SNIP>

City is not just a function of population - It’s can also be a political appointment/status? - e.g. Charters Towers and Redcliffe are cities : https://en.wikipedia.org/wiki/List_of_cities_in_Australia



Yes there is an 'official designation system' ... subject to political pressure and separate rules for each state.
I think the best guide we have is the population, certainly I think it is much better than the officially given 'status'.

----------------------
I did leave out of the original post that the ABS data may include more 'cities' with populations over 10,000 than the present OSM data base contains ... yet to sort that out.



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? An oldie but no past conclusion.

Alex Sims
I’ve had an involvement in this discussion in the past and wonder if a way forward might be to include an adjusting factor for remoteness.


which shows the Australian Remoteness Index this suggests that we could define town, hamlet, etc according to population but then adjust the population limits downward for remote areas.

The other point I’d make (as I did some time ago) is that the labels are “British English” labels and form a hierarchy where the names make sense in the UK but shouldn’t be taken as a slight against any area. They are merely a series of words that define the level of population centre. 

Looking at http://wiki.openstreetmap.org/wiki/Key:place#Populated_settlements.2C_urban_and_rural this seems to support and adjustment based on remoteness in the Australian context.

Alex

On 4 May 2016, at 8:11 AM, Warin <[hidden email]> wrote:

On 4/05/2016 12:50 AM, Christopher Barham wrote:

On 03 May 2016, at 14:22, Warin <[hidden email][hidden email]> wrote:


<SNIP>
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.
</SNIP>

City is not just a function of population - It’s can also be a political appointment/status? - e.g. Charters Towers and Redcliffe are cities : https://en.wikipedia.org/wiki/List_of_cities_in_Australia



Yes there is an 'official designation system' ... subject to political pressure and separate rules for each state.
I think the best guide we have is the population, certainly I think it is much better than the officially given 'status'.

----------------------
I did leave out of the original post that the ABS data may include more 'cities' with populations over 10,000 than the present OSM data base contains ... yet to sort that out.


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? An oldie but no past conclusion.

Ian Sergeant-2
I'd make two comments.

1. Any attempt to make something render on sparse parts of the map, is
a rendering issue.  Any renderer is free to pre-process the data based
on a population and remoteness algorithm if they wish.

2. Personally, I make anything a town if it has services.  If it has a
pub, a take-away, a supermarket, a post-office, and a fuel station,
then it's a town.  I save hamlet for a population grouping without any
services, and a locality for a place where there is essentially no
population clustering.  This is a natural skew towards remoter
destinations becoming towns, because they are service towns for
surrounding areas, rather than necessarily having large populations
themselves.

Ian.



On 4 May 2016 at 18:28, Alex Sims <[hidden email]> wrote:

> I’ve had an involvement in this discussion in the past and wonder if a way
> forward might be to include an adjusting factor for remoteness.
>
> If you have a look at the map at
> http://www.abs.gov.au/websitedbs/d3310114.nsf/home/remoteness+structure
>
> which shows the Australian Remoteness Index this suggests that we could
> define town, hamlet, etc according to population but then adjust the
> population limits downward for remote areas.
>
> The other point I’d make (as I did some time ago) is that the labels are
> “British English” labels and form a hierarchy where the names make sense in
> the UK but shouldn’t be taken as a slight against any area. They are merely
> a series of words that define the level of population centre.
>
> Looking at
> http://wiki.openstreetmap.org/wiki/Key:place#Populated_settlements.2C_urban_and_rural
> this seems to support and adjustment based on remoteness in the Australian
> context.
>
> Alex
>
> On 4 May 2016, at 8:11 AM, Warin <[hidden email]> wrote:
>
> On 4/05/2016 12:50 AM, Christopher Barham wrote:
>
>
> On 03 May 2016, at 14:22, Warin <[hidden email]> wrote:
>
>
> <SNIP>
>
> Why judge on the population?
>
> Larger populations get more services - Police, Medical, Education ... they
> go hand in hand.
>
> Populations are usually stated - on the entry signs to towns, villages ..
> and collected by the ABS. So verifiable and accessible.
>
> Yes they do change .. but not by vast amounts quickly.
>
> Usually the relationship between population centres remains fairly static ..
> if one grows so do the surrounding ones.
>
> Much easier to quickly asses and correctly tag this way. So it satisfies the
> KISS principle.
>
> </SNIP>
>
> City is not just a function of population - It’s can also be a political
> appointment/status? - e.g. Charters Towers and Redcliffe are cities :
> https://en.wikipedia.org/wiki/List_of_cities_in_Australia
>
>
>
> Yes there is an 'official designation system' ... subject to political
> pressure and separate rules for each state.
> I think the best guide we have is the population, certainly I think it is
> much better than the officially given 'status'.
>
> ----------------------
> I did leave out of the original post that the ABS data may include more
> 'cities' with populations over 10,000 than the present OSM data base
> contains ... yet to sort that out.
>
>
> _______________________________________________
> Talk-au mailing list
> [hidden email]
> https://lists.openstreetmap.org/listinfo/talk-au
>
>
>
> _______________________________________________
> Talk-au mailing list
> [hidden email]
> https://lists.openstreetmap.org/listinfo/talk-au
>

_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? An oldie but no past conclusion.

Simon Slater
In reply to this post by Alex Sims
On Wed, 4 May 2016 05:58:27 PM Alex Sims wrote:
> The other point I’d make (as I did some time ago) is that the labels are
> “British English” labels and form a hierarchy where the names make sense in
> the UK but shouldn’t be taken as a slight against any area. They are merely
> a series of words that define the level of population centre.

Looking at the end of this post:
https://lists.openstreetmap.org/pipermail/talk-au/2008-December/001089.html
made me think of our own experience with small towns.

In 1988 we moved to Kerang, Vic which had a population of 5,500, 5 pubs, 1
small supermarket, 1 large supermarket with bottleshop, 8 churches and little
industry.  However, Kerang supported a regional farming population of 20,000.  
When we moved 50 miles up the road 10 years later, the population was 4,500.  
10 years later 1 pub burned down, 5 years later so did another.  Now the
population is below 4000 I think, but the regional population serviced is
still about the same and there is more industry in the town.

I assume computerization and mechanization means the increase in industry
without population increase.  Also with amalgamation of farms, many houses are
now available for those who work in town, so these would not be counted in the
town stats.

My thought was to look at the amenities etc listed for a place within OSM
itself for use as a guide to classification.  Would this be a purely subjective
process ie looking at the map, or can this type of data be easily queried from
the database for a more objective approach?

In the latter case, weights could be applied to different amenities,
combination with other sources eg remoteness index, etc ...

The caveat here is that the more amenities mapped correlate with activity /
interest in that location.
--
Regards
Simon Slater

Registered Linux User #463789
http://linuxcounter.net 


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place by remoteness

Warin
In reply to this post by Alex Sims
Remoteness .. nice!
It is based on population density .. the same argument I make for lowering the population barriers for city/town/village for Australia. So, yes, I do like it.
How far to take the 'remoteness' effect on the population barriers to?
If the area has very little population then 1 person could be defined as a city? NO, certain things are expected in a city .. certainly more than 1 person!
So there are limits as to how far to go in this direction.

Would need to revert to
city>100,000>town>10,000>village>200>hamlet>100
for 'Major cities' and 'Inner regional' areas -
as judged by the 'remoteness' thing as I can see no reason not to use the world wide population points here as the population densities are similar?
These areas are in close proximity and would be similar around the world so the chosen population points should be suitable.

The 'Outer Regional' areas ... about half the population density so
city>50,000>town>5,000>village>100>hamlet>50
 The 'Remote' areas ... about half the population density so
city>25,000>town>2,500>village>50>hamlet>25
The 'Very Remote' areas ... about half the population density so
city>12,500>town>2,500>village>50>hamlet>25

Err Winton would be come a village .. Longreach becomes a town... would that be acceptable?
I think that works for my perception of those places.

It will add to the complexity but be justifiable technically. Is it worth the added complexity?
 
On 4/05/2016 6:28 PM, Alex Sims wrote:
I’ve had an involvement in this discussion in the past and wonder if a way forward might be to include an adjusting factor for remoteness.


which shows the Australian Remoteness Index this suggests that we could define town, hamlet, etc according to population but then adjust the population limits downward for remote areas.

The other point I’d make (as I did some time ago) is that the labels are “British English” labels and form a hierarchy where the names make sense in the UK but shouldn’t be taken as a slight against any area. They are merely a series of words that define the level of population centre. 

Looking at http://wiki.openstreetmap.org/wiki/Key:place#Populated_settlements.2C_urban_and_rural this seems to support and adjustment based on remoteness in the Australian context.

Alex

On 4 May 2016, at 8:11 AM, Warin <[hidden email]> wrote:

On 4/05/2016 12:50 AM, Christopher Barham wrote:

On 03 May 2016, at 14:22, Warin <[hidden email]> wrote:


<SNIP>
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.
</SNIP>

City is not just a function of population - It’s can also be a political appointment/status? - e.g. Charters Towers and Redcliffe are cities : https://en.wikipedia.org/wiki/List_of_cities_in_Australia



Yes there is an 'official designation system' ... subject to political pressure and separate rules for each state.
I think the best guide we have is the population, certainly I think it is much better than the officially given 'status'.

----------------------
I did leave out of the original post that the ABS data may include more 'cities' with populations over 10,000 than the present OSM data base contains ... yet to sort that out.


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place by remoteness

Ben Kelley-4

Hi.

The remoteness doesn't need to change the definition of the place (e.g. make a hamlet a town) but rather only change how it is rendered.

A very remote track might show, as might a remote hamlet.

I agree this might be difficult to implement in the renderer.

  - Ben.

--
Ben Kelley
[hidden email]
Sent from my Windows XP PC

On 5 May 2016 10:26, "Warin" <[hidden email]> wrote:
Remoteness .. nice!
It is based on population density .. the same argument I make for lowering the population barriers for city/town/village for Australia. So, yes, I do like it.
How far to take the 'remoteness' effect on the population barriers to?
If the area has very little population then 1 person could be defined as a city? NO, certain things are expected in a city .. certainly more than 1 person!
So there are limits as to how far to go in this direction.

Would need to revert to
city>100,000>town>10,000>village>200>hamlet>100
for 'Major cities' and 'Inner regional' areas -
as judged by the 'remoteness' thing as I can see no reason not to use the world wide population points here as the population densities are similar?
These areas are in close proximity and would be similar around the world so the chosen population points should be suitable.

The 'Outer Regional' areas ... about half the population density so
city>50,000>town>5,000>village>100>hamlet>50
 The 'Remote' areas ... about half the population density so
city>25,000>town>2,500>village>50>hamlet>25
The 'Very Remote' areas ... about half the population density so
city>12,500>town>2,500>village>50>hamlet>25

Err Winton would be come a village .. Longreach becomes a town... would that be acceptable?
I think that works for my perception of those places.

It will add to the complexity but be justifiable technically. Is it worth the added complexity?
 
On 4/05/2016 6:28 PM, Alex Sims wrote:
I’ve had an involvement in this discussion in the past and wonder if a way forward might be to include an adjusting factor for remoteness.


which shows the Australian Remoteness Index this suggests that we could define town, hamlet, etc according to population but then adjust the population limits downward for remote areas.

The other point I’d make (as I did some time ago) is that the labels are “British English” labels and form a hierarchy where the names make sense in the UK but shouldn’t be taken as a slight against any area. They are merely a series of words that define the level of population centre. 

Looking at http://wiki.openstreetmap.org/wiki/Key:place#Populated_settlements.2C_urban_and_rural this seems to support and adjustment based on remoteness in the Australian context.

Alex

On 4 May 2016, at 8:11 AM, Warin <[hidden email]> wrote:

On 4/05/2016 12:50 AM, Christopher Barham wrote:

On 03 May 2016, at 14:22, Warin <[hidden email]> wrote:


<SNIP>
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.
</SNIP>

City is not just a function of population - It’s can also be a political appointment/status? - e.g. Charters Towers and Redcliffe are cities : https://en.wikipedia.org/wiki/List_of_cities_in_Australia



Yes there is an 'official designation system' ... subject to political pressure and separate rules for each state.
I think the best guide we have is the population, certainly I think it is much better than the officially given 'status'.

----------------------
I did leave out of the original post that the ABS data may include more 'cities' with populations over 10,000 than the present OSM data base contains ... yet to sort that out.


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place now remoteness rendering

Warin
On 5/05/2016 10:35 AM, Ben Kelley wrote:

Hi.

The remoteness doesn't need to change the definition of the place (e.g. make a hamlet a town) but rather only change how it is rendered.

A very remote track might show, as might a remote hamlet.

I agree this might be difficult to implement in the renderer.


Then I think that is a separate issue to tagging.
I have modified the subject to reflect the divergence of subject.

Rendering where a large area is blank (or at least not 'full') I think the render can go to the next layer and render that .. repeat until there is enough to display.
This would fill the map with data - making the map more usefull.

  - Ben.

--
Ben Kelley
[hidden email]
Sent from my Windows XP PC

On 5 May 2016 10:26, "Warin" <[hidden email]> wrote:
Remoteness .. nice!
It is based on population density .. the same argument I make for lowering the population barriers for city/town/village for Australia. So, yes, I do like it.
How far to take the 'remoteness' effect on the population barriers to?
If the area has very little population then 1 person could be defined as a city? NO, certain things are expected in a city .. certainly more than 1 person!
So there are limits as to how far to go in this direction.

Would need to revert to
city>100,000>town>10,000>village>200>hamlet>100
for 'Major cities' and 'Inner regional' areas -
as judged by the 'remoteness' thing as I can see no reason not to use the world wide population points here as the population densities are similar?
These areas are in close proximity and would be similar around the world so the chosen population points should be suitable.

The 'Outer Regional' areas ... about half the population density so
city>50,000>town>5,000>village>100>hamlet>50
 The 'Remote' areas ... about half the population density so
city>25,000>town>2,500>village>50>hamlet>25
The 'Very Remote' areas ... about half the population density so
city>12,500>town>2,500>village>50>hamlet>25

Err Winton would be come a village .. Longreach becomes a town... would that be acceptable?
I think that works for my perception of those places.

It will add to the complexity but be justifiable technically. Is it worth the added complexity?
 
On 4/05/2016 6:28 PM, Alex Sims wrote:
I’ve had an involvement in this discussion in the past and wonder if a way forward might be to include an adjusting factor for remoteness.


which shows the Australian Remoteness Index this suggests that we could define town, hamlet, etc according to population but then adjust the population limits downward for remote areas.

The other point I’d make (as I did some time ago) is that the labels are “British English” labels and form a hierarchy where the names make sense in the UK but shouldn’t be taken as a slight against any area. They are merely a series of words that define the level of population centre. 

Looking at http://wiki.openstreetmap.org/wiki/Key:place#Populated_settlements.2C_urban_and_rural this seems to support and adjustment based on remoteness in the Australian context.

Alex

On 4 May 2016, at 8:11 AM, Warin <[hidden email]> wrote:

On 4/05/2016 12:50 AM, Christopher Barham wrote:

On 03 May 2016, at 14:22, Warin <[hidden email]> wrote:


<SNIP>
Why judge on the population? 
Larger populations get more services - Police, Medical, Education ... they go hand in hand. 
Populations are usually stated - on the entry signs to towns, villages .. and collected by the ABS. So verifiable and accessible.
Yes they do change .. but not by vast amounts quickly. 
Usually the relationship between population centres remains fairly static .. if one grows so do the surrounding ones. 
Much easier to quickly asses and correctly tag this way. So it satisfies the KISS principle.
</SNIP>

City is not just a function of population - It’s can also be a political appointment/status? - e.g. Charters Towers and Redcliffe are cities : https://en.wikipedia.org/wiki/List_of_cities_in_Australia



Yes there is an 'official designation system' ... subject to political pressure and separate rules for each state.
I think the best guide we have is the population, certainly I think it is much better than the officially given 'status'.

----------------------
I did leave out of the original post that the ABS data may include more 'cities' with populations over 10,000 than the present OSM data base contains ... yet to sort that out.


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? An oldie but no past conclusion.

Simon Slater
In reply to this post by Ian Sergeant-2
On Thu, 5 May 2016 10:10:35 AM Ian Sergeant wrote:

> 1. Any attempt to make something render on sparse parts of the map, is
> a rendering issue.  Any renderer is free to pre-process the data based
> on a population and remoteness algorithm if they wish.
>
> 2. Personally, I make anything a town if it has services.  If it has a
> pub, a take-away, a supermarket, a post-office, and a fuel station,
> then it's a town.  I save hamlet for a population grouping without any
> services, and a locality for a place where there is essentially no
> population clustering.  This is a natural skew towards remoter
> destinations becoming towns, because they are service towns for
> surrounding areas, rather than necessarily having large populations
> themselves.

Post offices may be a good guide.  25 years ago there were at least 4 post offices
between here (Swan Hill) and Kerang.  Now there is only one at Lake Boga, but
all the other post codes are still in place, mail routing through either
Kerang or Swan Hill.

The other places, Lake Charm, Mystic Park and Tresco, have a small group of
houses, Lk Charm has a shop and caravan park, Mystic has a pub.

The passenger train doesn't stop at any of these stations anymore either, only
the grain trains.
--
Regards
Simon Slater

Registered Linux User #463789
http://linuxcounter.net 


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: ����������������������������������������������������(

Warin
On 5/6/2016 9:51 AM, Simon Slater wrote:
On Thu, 5 May 2016 10:10:35 AM Ian Sergeant wrote:
1. Any attempt to make something render on sparse parts of the map, is
a rendering issue.  Any renderer is free to pre-process the data based
on a population and remoteness algorithm if they wish.

2. Personally, I make anything a town if it has services.  If it has a
pub, a take-away, a supermarket, a post-office, and a fuel station,
then it's a town.  I save hamlet for a population grouping without any
services, and a locality for a place where there is essentially no
population clustering.  This is a natural skew towards remoter
destinations becoming towns, because they are service towns for
surrounding areas, rather than necessarily having large populations
themselves.
Post offices may be a good guide.  25 years ago there were at least 4 post offices 
between here (Swan Hill) and Kerang.  Now there is only one at Lake Boga, but 
all the other post codes are still in place, mail routing through either 
Kerang or Swan Hill.

Australia post has;
  • Post Office (PO) and
  • Local Post Office (LPO)
The LPO is usually smaller and within another business ..usually a local convenience store. 

    
The ABS has this 
http://www.abs.gov.au/AUSSTATS/abs@.nsf/2f762f95845417aeca25706c00834efa/1d90c1ef4ac928d5ca2570ec0018e4f7!OpenDocument
"Identifying towns

In this review small towns have been defined as population centres with between 1,000 and 19,999 people. 
Towns might ideally be distinguished from cities and from smaller rural communities according to functional criteria, such as the 
presence or absence of various educational, medical, recreational and retail services, together perhaps with administrative criteria such as 
whether or not a city or town council operated from within the town. 
While such conceptual distinctions might be made, it is difficult to put such definitions into practice. The above population size was therefore
 considered the most suitable alternative which would generally encompass these criteria."

I tend to concur with this - simplest to implement and verify. 

I do note the 'medical' services that ABS have for identifying towns etc, that may be a usefull criteria in addition to number of pubs, petrol stations etc.  




_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: ����������������������������������������������������(

Warin
On 5/13/2016 11:36 AM, Warin wrote:
On 5/6/2016 9:51 AM, Simon Slater wrote:
On Thu, 5 May 2016 10:10:35 AM Ian Sergeant wrote:
1. Any attempt to make something render on sparse parts of the map, is
a rendering issue.  Any renderer is free to pre-process the data based
on a population and remoteness algorithm if they wish.

2. Personally, I make anything a town if it has services.  If it has a
pub, a take-away, a supermarket, a post-office, and a fuel station,
then it's a town.  I save hamlet for a population grouping without any
services, and a locality for a place where there is essentially no
population clustering.  This is a natural skew towards remoter
destinations becoming towns, because they are service towns for
surrounding areas, rather than necessarily having large populations
themselves.
Post offices may be a good guide.  25 years ago there were at least 4 post offices 
between here (Swan Hill) and Kerang.  Now there is only one at Lake Boga, but 
all the other post codes are still in place, mail routing through either 
Kerang or Swan Hill.

Australia post has;
  • Post Office (PO) and
  • Local Post Office (LPO)
The LPO is usually smaller and within another business ..usually a local convenience store. 
The ABS has this 
http://www.abs.gov.au/AUSSTATS/abs@.nsf/2f762f95845417aeca25706c00834efa/1d90c1ef4ac928d5ca2570ec0018e4f7!OpenDocument
"Identifying towns

In this review small towns have been defined as population centres with between 1,000 and 19,999 people. 
Towns might ideally be distinguished from cities and from smaller rural communities according to functional criteria, such as the 
presence or absence of various educational, medical, recreational and retail services, together perhaps with administrative criteria such as 
whether or not a city or town council operated from within the town. 
While such conceptual distinctions might be made, it is difficult to put such definitions into practice. The above population size was therefore
 considered the most suitable alternative which would generally encompass these criteria."

I tend to concur with this - simplest to implement and verify. 

I do note the 'medical' services that ABS have for identifying towns etc, that may be a usefull criteria in addition to number of pubs, petrol stations etc.  


I have gotten some 1,400 'towns from the OSM data base .. many of these have no population given, but from those that do;
Penrith		178465
Bunbury		64385
Maitland	61431
Palmerston	46618
Melton		45624
Port Macquarie	41723
Sunbury		33062
Pakenham	32911
Nowra		32556
Albany		30656
Devonport	29051
Goulburn	21484
Busselton	21407
Ocean Grove	16093
Bacchus Marsh	14913
Port Hedland	13772
Torquay		13339
Coolum Beach	13154
Broome		12766
Batemans Bay	12000
Lara		11192
Drysdale	10927

Compare this to the 'cities';
 
Charters Towers	8,234
Charleville	4,700
Caloundra	3,550
Winton		1,337 



Winton and Charleville 'cities' when Broome is not? No .. sorry that makes no sense, even when taking into account 'remoteness' and services. 

And the other end of 'towns';

Marble Bar	194
Coral Bay	190
Coorow		176
Guilderton	146
Marvel Loch	98
Popanyinning	87
Betoota		0
Ooldea		0

Ooldea used to have a hermit, he could still be kicking ... the others there were railway workers with homes elsewhere.
Betoota used to have 1, but he died. 
I don't think anyone could call these 'towns'! Even on a 'services' scale. 

There is a clear disparity here. 


I hope to get all the 'towns' populations that are missing entered from the ABS census 2011, and then look again at these 'towns'. 
I have entered all the 'cities' population data that was missing, so that bit is done. 

I am yet to get the villages list, if I do! 

----------------------- 
As a reminder of what I first proposed
OSM wiki presently has by population
city>100,000>town>10,000>village>200>hamlet>100
      
I think for Australia;
city>10,000>town>1,000>village>100>hamlet>10

 



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: ����������������������������������������������������(

cleary
 
I agree that there is a need to improve our classification of places. However I think that taking population as the sole criterion will create more discrepancies than we have already.
 
For example, I live in a Sydney suburb that has a population greater than the gazetted "state suburb" of Sydney (roughly the CBD area). If we adopted a strictly population-based criterion, my suburb and many others with more than 10,000 people would be "towns" in OSM and Sydney CBD be a "town". My suburb has about the same population as the rural city of Griffith, NSW. I think Griffith is a city but my suburb is not.
 
The figure extracted from the OSM population data shows Nowra with over 32,000 population. On the ABS website, the ABS census data for 2011 shows the "state suburb" of Nowra having 9,200 people .. just a village. If we add the adjacent areas of North Nowra, South Nowra, West Nowra and Bomaderry, then ABS shows the "Nowra/Bomaderry Urban Centre" which includes the areas around Nowra as having approx. 28,000 people. So do we show Bomaderry as part of greater Nowra, or is it a village/town in its own right?  Does North Nowra exist in its own right as a suburb/village or is it just part of the larger Nowra?
 
There is also the issue of population density. 500 people living in a couple of square kilometres might be a village. Another 500 people living in a gazetted location of hundreds of square kilometres could hardly be said to be living in a village even though their location has an equivalent population.
 
I won't keep going on and on, but there are many questions thrown up by relying on population alone as the criterion for determining if a place is a city or town or whatever. I think it has to be a sort of "common sense" decision taking population into account but other factors as well.  But I do support the need to try to clarify our classifications and appreciate the difficulty in resolving the issue.
 
 
 
 
 
On Fri, May 13, 2016, at 07:11 PM, Warin wrote:
On 5/13/2016 11:36 AM, Warin wrote:
On 5/6/2016 9:51 AM, Simon Slater wrote:
On Thu, 5 May 2016 10:10:35 AM Ian Sergeant wrote:

1. Any attempt to make something render on sparse parts of the map, is
a rendering issue.  Any renderer is free to pre-process the data based
on a population and remoteness algorithm if they wish.

2. Personally, I make anything a town if it has services.  If it has a
pub, a take-away, a supermarket, a post-office, and a fuel station,
then it's a town.  I save hamlet for a population grouping without any
services, and a locality for a place where there is essentially no
population clustering.  This is a natural skew towards remoter
destinations becoming towns, because they are service towns for
surrounding areas, rather than necessarily having large populations
themselves.

Post offices may be a good guide.  25 years ago there were at least 4 post offices 
between here (Swan Hill) and Kerang.  Now there is only one at Lake Boga, but 
all the other post codes are still in place, mail routing through either 
Kerang or Swan Hill.
 
Australia post has;
  • Post Office (PO) and
  • Local Post Office (LPO)
The LPO is usually smaller and within another business ..usually a local convenience store. 

The ABS has this 

http://www.abs.gov.au/AUSSTATS/abs@.nsf/2f762f95845417aeca25706c00834efa/1d90c1ef4ac928d5ca2570ec0018e4f7!OpenDocument
"Identifying towns

In this review small towns have been defined as population centres with between 1,000 and 19,999 people. 
Towns might ideally be distinguished from cities and from smaller rural communities according to functional criteria, such as the 
presence or absence of various educational, medical, recreational and retail services, together perhaps with administrative criteria such as 
whether or not a city or town council operated from within the town. 
While such conceptual distinctions might be made, it is difficult to put such definitions into practice. The above population size was therefore
 considered the most suitable alternative which would generally encompass these criteria."


I tend to concur with this - simplest to implement and verify. 

I do note the 'medical' services that ABS have for identifying towns etc, that may be a usefull criteria in addition to number of pubs, petrol stations etc.  


I have gotten some 1,400 'towns from the OSM data base .. many of these have no population given, but from those that do;
Penrith		178465
Bunbury		64385
Maitland	61431
Palmerston	46618
Melton		45624
Port Macquarie	41723
Sunbury		33062
Pakenham	32911
Nowra		32556
Albany		30656
Devonport	29051
Goulburn	21484
Busselton	21407
Ocean Grove	16093
Bacchus Marsh	14913
Port Hedland	13772
Torquay		13339
Coolum Beach	13154
Broome		12766
Batemans Bay	12000
Lara		11192
Drysdale	10927

Compare this to the 'cities';
 
Charters Towers	8,234
Charleville	4,700
Caloundra	3,550
Winton		1,337 



Winton and Charleville 'cities' when Broome is not? No .. sorry that makes no sense, even when taking into account 'remoteness' and services. 

And the other end of 'towns';

Marble Bar	194
Coral Bay	190
Coorow		176
Guilderton	146
Marvel Loch	98
Popanyinning	87
Betoota		0
Ooldea		0

Ooldea used to have a hermit, he could still be kicking ... the others there were railway workers with homes elsewhere.
Betoota used to have 1, but he died. 
I don't think anyone could call these 'towns'! Even on a 'services' scale. 

There is a clear disparity here. 


I hope to get all the 'towns' populations that are missing entered from the ABS census 2011, and then look again at these 'towns'. 
I have entered all the 'cities' population data that was missing, so that bit is done. 

I am yet to get the villages list, if I do! 

----------------------- 
As a reminder of what I first proposed
OSM wiki presently has by population
city>100,000>town>10,000>village>200>hamlet>100
      
I think for Australia;
city>10,000>town>1,000>village>100>hamlet>10

 

 


_______________________________________________
Talk-au mailing list
 

_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place=? and oldie but a goodie.

Simon Slater
In reply to this post by Warin
On Fri, 13 May 2016 07:11:39 PM Warin wrote:
2 things:
First, I may have mussed up the threading here but the subject of Warin's last
thread looked like "ᅵᅵᅵᅵ...".  My e-mail client has been playing
silly-buggers lately, so was this the intended subject or just my client?

> I have gotten some 1,400 'towns from the OSM data base ..

Second, how does one extract data like this from the OSM database?  Is there a
wikipage or other guide?
--
Regards
Simon Slater

Registered Linux User #463789
http://linuxcounter.net 


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place town subject and Download

Warin
On 5/14/2016 11:30 AM, Simon Slater wrote:
> On Fri, 13 May 2016 07:11:39 PM Warin wrote:
> 2 things:
> First, I may have mussed up the threading here but the subject of Warin's last
> thread looked like "ᅵᅵᅵᅵ...".  My e-mail client has been playing
> silly-buggers lately, so was this the intended subject or just my client?

Yer... I think that is the use of ? and/or = characters that much up these email clients. Sorry about that, did not realise.

>
>> I have gotten some 1,400 'towns from the OSM data base ..
> Second, how does one extract data like this from the OSM database?  Is there a
> wikipage or other guide?

I am using JOSM -

file - Download from Overpass API
Build query ... enter 'place=town' in the box on the right then  click on the 'Build query' and that gets the Overpass query built blow your entry.
Then select the area .. if you make this too large it will time out. Download...

This gets you it into JOSM .. then save as to get a .osm file and I then use a small script to extract the name and population from that file into a .csv file.

I just followed my nose to get here.


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place equals what BY features

Warin
In reply to this post by cleary
On 5/13/2016 9:22 PM, cleary wrote:
 
I agree that there is a need to improve our classification of places. However I think that taking population as the sole criterion will create more discrepancies than we have already.

I think of it as a guide. In fact most of the OSM wiki to me is a guide.

 
For example, I live in a Sydney suburb that has a population greater than the gazetted "state suburb" of Sydney (roughly the CBD area). If we adopted a strictly population-based criterion, my suburb and many others with more than 10,000 people would be "towns" in OSM and Sydney CBD be a "town". My suburb has about the same population as the rural city of Griffith, NSW. I think Griffith is a city but my suburb is not.
Yep. I take your point. Closer to home is Penrith .. a city or a suburb of Sydney?
 
I won't keep going on and on, but there are many questions thrown up by relying on population alone as the criterion for determining if a place is a city or town or whatever. I think it has to be a sort of "common sense" decision taking population into account but other factors as well.  But I do support the need to try to clarify our classifications and appreciate the difficulty in resolving the issue.
 
 
A start on the classification by features? Warning .. draft only!

A city at a minimum has;

one hospital with emergency services
more than one police station
more than one public library
more than one secondary school
a university
more than one doctor's practice
more than one petrol station
more than one bank
more than one ATM
more than one Post Office

A town at a minimum has;

a hospital
a police station
a public library
a secondary school
a doctor's practice
a newsagent
a petrol station
a bank
a Post Office

A village at a minimum has;

a convenience store



 
 
 
On Fri, May 13, 2016, at 07:11 PM, Warin wrote:
On 5/13/2016 11:36 AM, Warin wrote:
On 5/6/2016 9:51 AM, Simon Slater wrote:
On Thu, 5 May 2016 10:10:35 AM Ian Sergeant wrote:

1. Any attempt to make something render on sparse parts of the map, is
a rendering issue.  Any renderer is free to pre-process the data based
on a population and remoteness algorithm if they wish.

2. Personally, I make anything a town if it has services.  If it has a
pub, a take-away, a supermarket, a post-office, and a fuel station,
then it's a town.  I save hamlet for a population grouping without any
services, and a locality for a place where there is essentially no
population clustering.  This is a natural skew towards remoter
destinations becoming towns, because they are service towns for
surrounding areas, rather than necessarily having large populations
themselves.

Post offices may be a good guide.  25 years ago there were at least 4 post offices 
between here (Swan Hill) and Kerang.  Now there is only one at Lake Boga, but 
all the other post codes are still in place, mail routing through either 
Kerang or Swan Hill.
 
Australia post has;
  • Post Office (PO) and
    
  • Local Post Office (LPO)
    
The LPO is usually smaller and within another business ..usually a local convenience store. 

The ABS has this 

http://www.abs.gov.au/AUSSTATS/abs@.nsf/2f762f95845417aeca25706c00834efa/1d90c1ef4ac928d5ca2570ec0018e4f7!OpenDocument
"Identifying towns

In this review small towns have been defined as population centres with between 1,000 and 19,999 people. 
Towns might ideally be distinguished from cities and from smaller rural communities according to functional criteria, such as the 
presence or absence of various educational, medical, recreational and retail services, together perhaps with administrative criteria such as 
whether or not a city or town council operated from within the town. 
While such conceptual distinctions might be made, it is difficult to put such definitions into practice. The above population size was therefore
 considered the most suitable alternative which would generally encompass these criteria."


I tend to concur with this - simplest to implement and verify. 

I do note the 'medical' services that ABS have for identifying towns etc, that may be a usefull criteria in addition to number of pubs, petrol stations etc.  


I have gotten some 1,400 'towns from the OSM data base .. many of these have no population given, but from those that do;
Penrith		178465
Bunbury		64385
Maitland	61431
Palmerston	46618
Melton		45624
Port Macquarie	41723
Sunbury		33062
Pakenham	32911
Nowra		32556
Albany		30656
Devonport	29051
Goulburn	21484
Busselton	21407
Ocean Grove	16093
Bacchus Marsh	14913
Port Hedland	13772
Torquay		13339
Coolum Beach	13154
Broome		12766
Batemans Bay	12000
Lara		11192
Drysdale	10927

Compare this to the 'cities';
 
Charters Towers	8,234
Charleville	4,700
Caloundra	3,550
Winton		1,337 



Winton and Charleville 'cities' when Broome is not? No .. sorry that makes no sense, even when taking into account 'remoteness' and services. 

And the other end of 'towns';

Marble Bar	194
Coral Bay	190
Coorow		176
Guilderton	146
Marvel Loch	98
Popanyinning	87
Betoota		0
Ooldea		0

Ooldea used to have a hermit, he could still be kicking ... the others there were railway workers with homes elsewhere.
Betoota used to have 1, but he died. 
I don't think anyone could call these 'towns'! Even on a 'services' scale. 

There is a clear disparity here. 


I hope to get all the 'towns' populations that are missing entered from the ABS census 2011, and then look again at these 'towns'. 
I have entered all the 'cities' population data that was missing, so that bit is done. 

I am yet to get the villages list, if I do! 

----------------------- 
As a reminder of what I first proposed
OSM wiki presently has by population
city>100,000>town>10,000>village>200>hamlet>100
      
I think for Australia;
city>10,000>town>1,000>village>100>hamlet>10

 

 


_______________________________________________
Talk-au mailing list
 


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au



_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place town subject and Download

Marc Gemis
In reply to this post by Warin
with this query, you immediately have a CSV file as output.

Use this  http://overpass-turbo.eu/s/gdp

or copy the query below to http://overpass-turbo.eu/



[out:csv (name, population)][timeout:25];
// gather results
(
  // query part for: “place=town”
  node["place"="town"]({{bbox}});
  way["place"="town"]({{bbox}});
  relation["place"="town"]({{bbox}});
);
// print results
out body;
>;
out skel qt;

see http://wiki.openstreetmap.org/wiki/Overpass_API/Overpass_QL#Output_Format_.28out.29
for more options on the CSV output

You can make the timeout longer in case you want to cover a larger area.

regards

m

On Sat, May 14, 2016 at 4:48 AM, Warin <[hidden email]> wrote:

> On 5/14/2016 11:30 AM, Simon Slater wrote:
>>
>> On Fri, 13 May 2016 07:11:39 PM Warin wrote:
>> 2 things:
>> First, I may have mussed up the threading here but the subject of Warin's
>> last
>> thread looked like "ᅵᅵᅵᅵ...".  My e-mail client has been playing
>> silly-buggers lately, so was this the intended subject or just my client?
>
>
> Yer... I think that is the use of ? and/or = characters that much up these
> email clients. Sorry about that, did not realise.
>
>>
>>> I have gotten some 1,400 'towns from the OSM data base ..
>>
>> Second, how does one extract data like this from the OSM database?  Is
>> there a
>> wikipage or other guide?
>
>
> I am using JOSM -
>
> file - Download from Overpass API
> Build query ... enter 'place=town' in the box on the right then  click on
> the 'Build query' and that gets the Overpass query built blow your entry.
> Then select the area .. if you make this too large it will time out.
> Download...
>
> This gets you it into JOSM .. then save as to get a .osm file and I then use
> a small script to extract the name and population from that file into a .csv
> file.
>
> I just followed my nose to get here.
>
>
> _______________________________________________
> Talk-au mailing list
> [hidden email]
> https://lists.openstreetmap.org/listinfo/talk-au

_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place town subject and Download

Simon Slater
On Sat, 14 May 2016 05:34:46 PM Marc Gemis wrote:
> with this query, you immediately have a CSV file as output.
Thanks, I'll have a play with this today.  I have JOSM installed, but have
only had a cursory glance at it yet.
--
Regards
Simon Slater

Registered Linux User #463789
http://linuxcounter.net 


_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
Reply | Threaded
Open this post in threaded view
|

Re: place equals what BY features

cleary
In reply to this post by Warin
 
I think using services is more useful.  The classification "city" is a problem as it has multiple meanings (it can mean the inner city central business district or it can mean the larger city local government area or it can mean the very much larger area of surrounding communities) but other classifications can be straightforward.  In my travels (more extensive in NSW and Queensland than elsewhere), I have found local government areas to be a helpful guide to thinking about places and services. Towns generally provide services for people within their own local government area while cities provide services for people from greater distances including other local government areas.
 
 
My preliminary thinking is
 
 
Location : named place without residents.
 
 
Hamlet : place in which people reside but it does not necessarily provide any services to residents. Residents may be in a concentrated area or dispersed over a wider rural area.
 
 
Village : provides minimal services such as convenience store, fuel, hotel, primary school or multiple of these ... but not necessarily all of these. A village usually doesn't have a doctor but may provide some outpatient health services and/or an emergency ambulance to take people to a town for medical treatment but is unlikely to have in-patient health care.
 
 
Town : provides services not only for immediate residents but also for surrounding places, usually within the same local government area. Services might include health service/hospital, secondary school, specialty shops such as clothing store, hardware store, electrical store, major national or intenational fast-food store, local government offices, library ... most but not necessarily all of these. Just one of these is probably not enough to make a town - needs a few or even most of these. The presence of a high school (or combined primary/secondary) is a very good single indicator as this is a service for a significant surrounding area and cannot be sustained in smaller centres.  A hospital providing in-patient care is another good indicator.
 
 
City :  provides even more and higher level services and is a major centre for surrounding areas. The high level of services will attract people from surrounding local government areas.
 
Smaller cities are generally contained within their own local government areas but may provide services for residents from surrounding local government areas. The city central business district (CBD) can be classified as a "city" while surrounding areas within the same local government area are "suburbs" although it is a matter for the local government body to decide whether to subdivide into suburbs. For example, in NSW, the rural cities of Wagga Wagga and Griffith have central CBDs and the surrounding areas within the city local government area are officially classified as suburbs. In contrast, the whole of the Dubbo residential area is part of Dubbo and there neighbourhoods but no official suburbs. It would still be a city even though it has no offical suburbs. I think a key indicator is that cities usually provide services for people from surrounding local government areas as well as local residents. Services in cities are similar to towns but on a larger scale and higher level. For example the court house would house sittings of the District Court, education probably extends to a university campus, health and hospital services include specialist/referral services. A town might have a small number of professionals such as solicitors/doctors etc but a city would have more. A city has many more shops and offices than a town.
 
 
Then there are the larger cities such as capital cities - they can have cities within cities. There is central Sydney city (the CBD), then there is the City of Sydney local government area, but then there is the vast area commonly known as "greater Sydney" .... leading to some confusion when deciding what constitutes a city.  Within "greater Sydney" there are smaller cities with their own satellite suburbs. Parramatta, Blacktown, and Penrith are such smaller cities within greater Sydney. Then there is Chatswood ... the local government area is known as "Willoughby City Council" but Chatswood is the actual city CBD (I haven't though this through but I think Chatswood is the city location in this case as that is the place where the main services are - the name of the local government body is less relevant.) Then there can be "in-between" cities. Newcastle in NSW is not the state capital but it rivals smaller state/territory capitals in size and services. It has other local goverment areas within the "greater Newcastle" area making it more like a larger city than a smaller one.
 
I have no difficulty in naming Sydney, Chatswood, Penrith, Blacktown, Parramatta, Wagga Wagga, Griffith, Dubbo and Newcastle as cities and the same would apply to places elsewhere in Australia providing similar levels of services to residents of their own and surrounding local government areas.  
 
 
As a further thought, I am a little averse to using post offices as a guide. While would have been useful in the past, they are being closed and merged at a rapid rate and may almost disappear in the future.
 
Do my thoughts help at all?
 
 
 
 
On Sat, May 14, 2016, at 02:30 PM, Warin wrote:
On 5/13/2016 9:22 PM, cleary wrote:
 
I agree that there is a need to improve our classification of places. However I think that taking population as the sole criterion will create more discrepancies than we have already.
 
I think of it as a guide. In fact most of the OSM wiki to me is a guide.

 
For example, I live in a Sydney suburb that has a population greater than the gazetted "state suburb" of Sydney (roughly the CBD area). If we adopted a strictly population-based criterion, my suburb and many others with more than 10,000 people would be "towns" in OSM and Sydney CBD be a "town". My suburb has about the same population as the rural city of Griffith, NSW. I think Griffith is a city but my suburb is not.
Yep. I take your point. Closer to home is Penrith .. a city or a suburb of Sydney?
 
 
I won't keep going on and on, but there are many questions thrown up by relying on population alone as the criterion for determining if a place is a city or town or whatever. I think it has to be a sort of "common sense" decision taking population into account but other factors as well.  But I do support the need to try to clarify our classifications and appreciate the difficulty in resolving the issue.
 
 
A start on the classification by features? Warning .. draft only!

A city at a minimum has;

one hospital with emergency services
more than one police station
more than one public library
more than one secondary school
a university
more than one doctor's practice
more than one petrol station
more than one bank
more than one ATM
more than one Post Office

A town at a minimum has;

a hospital
a police station
a public library
a secondary school
a doctor's practice
a newsagent
a petrol station
a bank
a Post Office

A village at a minimum has;

a convenience store
 
 
 
 
 
 
On Fri, May 13, 2016, at 07:11 PM, Warin wrote:
On 5/13/2016 11:36 AM, Warin wrote:
On 5/6/2016 9:51 AM, Simon Slater wrote:
On Thu, 5 May 2016 10:10:35 AM Ian Sergeant wrote:


1. Any attempt to make something render on sparse parts of the map, is
a rendering issue.  Any renderer is free to pre-process the data based
on a population and remoteness algorithm if they wish.

2. Personally, I make anything a town if it has services.  If it has a
pub, a take-away, a supermarket, a post-office, and a fuel station,
then it's a town.  I save hamlet for a population grouping without any
services, and a locality for a place where there is essentially no
population clustering.  This is a natural skew towards remoter
destinations becoming towns, because they are service towns for
surrounding areas, rather than necessarily having large populations
themselves.


Post offices may be a good guide.  25 years ago there were at least 4 post offices 
between here (Swan Hill) and Kerang.  Now there is only one at Lake Boga, but 
all the other post codes are still in place, mail routing through either 
Kerang or Swan Hill.

 
Australia post has;

  • Post Office (PO) and
    
  • Local Post Office (LPO)
    
The LPO is usually smaller and within another business ..usually a local convenience store. 


The ABS has this 


http://www.abs.gov.au/AUSSTATS/abs@.nsf/2f762f95845417aeca25706c00834efa/1d90c1ef4ac928d5ca2570ec0018e4f7!OpenDocument

"Identifying towns

In this review small towns have been defined as population centres with between 1,000 and 19,999 people. 
Towns might ideally be distinguished from cities and from smaller rural communities according to functional criteria, such as the 
presence or absence of various educational, medical, recreational and retail services, together perhaps with administrative criteria such as 
whether or not a city or town council operated from within the town. 
While such conceptual distinctions might be made, it is difficult to put such definitions into practice. The above population size was therefore
 considered the most suitable alternative which would generally encompass these criteria."




I tend to concur with this - simplest to implement and verify. 

I do note the 'medical' services that ABS have for identifying towns etc, that may be a usefull criteria in addition to number of pubs, petrol stations etc.  



I have gotten some 1,400 'towns from the OSM data base .. many of these have no population given, but from those that do;

Penrith		178465
Bunbury		64385
Maitland	61431
Palmerston	46618
Melton		45624
Port Macquarie	41723
Sunbury		33062
Pakenham	32911
Nowra		32556
Albany		30656
Devonport	29051
Goulburn	21484
Busselton	21407
Ocean Grove	16093
Bacchus Marsh	14913
Port Hedland	13772
Torquay		13339
Coolum Beach	13154
Broome		12766
Batemans Bay	12000
Lara		11192
Drysdale	10927

Compare this to the 'cities';
 
Charters Towers	8,234
Charleville	4,700
Caloundra	3,550
Winton		1,337 



Winton and Charleville 'cities' when Broome is not? No .. sorry that makes no sense, even when taking into account 'remoteness' and services. 

And the other end of 'towns';

Marble Bar	194
Coral Bay	190
Coorow		176
Guilderton	146
Marvel Loch	98
Popanyinning	87
Betoota		0
Ooldea		0

Ooldea used to have a hermit, he could still be kicking ... the others there were railway workers with homes elsewhere.
Betoota used to have 1, but he died. 
I don't think anyone could call these 'towns'! Even on a 'services' scale. 

There is a clear disparity here. 


I hope to get all the 'towns' populations that are missing entered from the ABS census 2011, and then look again at these 'towns'. 
I have entered all the 'cities' population data that was missing, so that bit is done. 

I am yet to get the villages list, if I do! 

----------------------- 
As a reminder of what I first proposed
OSM wiki presently has by population
city>100,000>town>10,000>village>200>hamlet>100
      
I think for Australia;
city>10,000>town>1,000>village>100>hamlet>10

 


 


_______________________________________________
Talk-au mailing list
 
 
 
_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au


_______________________________________________
Talk-au mailing list
 

_______________________________________________
Talk-au mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk-au
12