Defining city boundaries is a question that is not only dealt by population density threshold - one wishes that it would be that simple. With suburban sprawl the merging of rural and suburban and urban make it very difficult to say where a city stops.
Conurbation definition, aire urbaine, in France take into account how far apart are the building (less than 200m), how many people that are resident in an area are actually working (40% and more)in the conurbation which must have at least 5000 jobs - an urban pole. An urban area is made of several urban pole - i.e. polycentric city.
It is all very relational and not just an attribute like density threshold.
In France an aire urbaine (literally: "urban area") is roughly the equivalent of a US Metropolitan Statistical Area. The name is extremely confusing because in English an "urban area" is not the same thing as a "metropolitan area". In English an urban area is a contiguously built-up area, which is smaller than a metropolitan area. Despite its name, a French aire urbaine is indeed a metropolitan area.
http://fr.wikipedia.org/wiki/Aire_urbaine_%28France%29 give an idea of what it look like for Nantes.
A distribution is given so you can reflect on the best fitting function.
From: Rui Carvalho [
Sent: 03 June 2007 10:02
Subject: Re: What is a street?
It seems to me that a city can be defined by population density? It is no
longer a city below a certain population density threshold?
Surely we'd have to agree on the threshold value (which is an arbitrary
parameter), but at least there is only one arbitrary parameter to agree
on, so it doesn't seem too problematic to me?
now a street...
On Sat, 2 Jun 2007 19:20:39 +0100, Stephen Marshall <[log in to unmask]>
>I still think it all depends on the purpose of
>what you are analysing, and the conclusions you
>intend to draw from the analysis.
>You could define and analyse
>- the public street network (as defined legally)
>- this may include 'fat' named spaces such as
>squares - for example in UK traffic regulation
>orders would typically include a combination of
>the named street/square and the physical extent
>(e.g. running from a to b) where the entity so defined need not be linear;
>- all private networks that are publicly
>accessible (with all the ramps and yards and royal parks)
>- all the private sections that are not publicly accessible;
>- all named streets / addresses (including courtyards, housing estates,
>- all axial lines
>- all numbered roads, etc.
>...and measure properties of each of these
>netowrks individually and in combination and get
>different numbers, that will tell you different things.
>The problem of comparability of networks across
>cultures would seem no more or less difficult
>than the problem of comparing anything, for
>example, cities. (If comparing London Paris
>Brasilia, etc., how do you decide what to count as consituting the city?).
>(What is a city?)
>At 12:12 02/06/2007, you wrote:
>>We should not be interested in the LEGAL STATUS
>>of places but whether or not there is actual
>>PRIVATE PHYSICAL CONTROL over their access at
>>any time. If there is, sorry, they are out of
>>the analysis AS PUBLIC PLACES, no matter that
>>thousands of people love to interact WITHIN
>>shopping centres or any other kind of
>>"non-places" (Auge) - very typical of present day societies by the way.
>>If encounters in a particular society happen
>>predominantly in private spaces, this is indeed
>>a cultural trait that has to be taken into
>>account, but this is another matter altogether,
>>it has nothing to do with the study of the
>>structure of the public realm, and the study of the street belongs here.
>>Sorry, syntax again...
>>Frederico de Holanda
>>Cond. Vivendas Colorado 1, Mod. J, Casa 1
>>73070-015 Brasília DF
>>Fone / Phone: (0xx61) 34859641 / +556134859641
>>Celular / Mobile: (0xx61) 99861724 / +556199861724
>>----- Original Message -----
>>From: <mailto:[log in to unmask]>Rui Carvalho
>>To: <mailto:[log in to unmask]>[log in to unmask]
>>Sent: Saturday, June 02, 2007 6:36 AM
>>Subject: Re: What is a street?
>>On Fri, 1 Jun 2007 12:00:52 -0300, Frederico de
>>Holanda <<mailto:[log in to unmask]>[log in to unmask]>
>> >In making axial maps, we use a very simple and unambiguous procedure to
>>identify a STREET - and plot it in the axial map: it must be PUBLIC, i.e.
>>there are no obstacles (any sort of PHYSICAL control) preventing people
>>from accessing it from urban space at large.
>> >"Streets" in CLOSED condominiums are ignored and are NOT ploted in the
>> >axial map - no matter how large the condominiums are (and they can be
>> >quite large in Brazilian cities). And we did not need "semantics" to
>> >this is pure syntax... After all, concerning streets, we are studying
>> >structure of the PUBLIC OPEN URBAN SPACE, aren't we?...
>>Well I thought you were studying the space where social interaction can
>>happen (the social logic of ...?)... and in the UK that does not need to
>>be public? The parks in London, for example, aren't they owned by the
>>Royal family? Aren't they private? What about the square around Abbey in
>>Euston Rd? Great (private) place to eat your lunch if you work in the
>>area... BTW, I've had great social interaction on the ramp of Torrington
>>The point here is that the concept of public or private is cultural: what
>>is considered private in one country may not be private in another. So
>>looks like you DO need semantic information after all...
>>Try again ;)
>> >If we are to consider "private streets" I suggest we add to the axial
>>of London the axial map of the internal spaces of the British Museum, the
>>Tate Gallery, The Royal Festival Hall, the residential towers in the
>> >Frederico de Holanda
>> >Cond. Vivendas Colorado 1, Mod. J, Casa 1
>> >73070-015 Brasília DF
>> >Fone / Phone: (0xx61) 34859641 / +556134859641
>> >Celular / Mobile: (0xx61) 99861724 / +556199861724
>> > ----- Original Message -----
>> > From: Lucas Figueiredo
>> > To: <mailto:[log in to unmask]>[log in to unmask]
>> > Sent: Friday, June 01, 2007 9:17 AM
>> > Subject: Re: What is a street?
>> > On 01/06/07, Hoon Park
>> <<mailto:[log in to unmask]>[log in to unmask]> wrote:
>> > > So, if applying this 'rule' and rephrasing my question, can we say
>> > > little droop in the lower-tail of an observed degree distribution
>> > > reflect the incompleteness of the data in keeping a consistent
>> > > resolution? Or consequently, if continuity aggregation, or any
>> > > makes the power-law hold better, can we say it is therefore a
>> > > of representing a street network? Or else, is a 'street' perhaps
>> > > linear aggregation of spaces that entails a power-law degree
>> > > at the higher-order level?
>> > It can reflect both things. The experiment is not independent from
>> > scientist. The lower-tails may be a problem with the data or just the
>> > real phenomenon, why discard this hyphotesis? On the other hand, the
>> > scientist may be well 'tweaking' the aggregation process, not to
>> > 'observe' the phenomenon, but to 'create' it.
>> > Apparently, as scientists does not care anymore to explain what they
>> > are assuming before the experiments, being allowed even to ignore the
>> > existence of whole fields such as space syntax, it is difficult to
>> > judge what is the phenomenon and what is simple play with different
>> > methods that will generate the phenomenon.
>> > Are we observing things or simulating things?
>> > Are we starting from questions or from answers?
>> > Best Regards,
>> > Lucas Figueiredo
>> > Mindwalk
>> > <http://www.mindwalk.com.br>http://www.mindwalk.com.br
>> > __________ Informação do NOD32 2304 (20070601) __________
>> > Esta mensagem foi verificada pelo NOD32 Sistema Antivírus
>> > <http://www.nod32.com.br>http://www.nod32.com.br
>>__________ Informação do NOD32 2305 (20070601) __________
>>Esta mensagem foi verificada pelo NOD32 Sistema Antivírus