[OSM-talk] Mapnik: strange prioritisation/appearance of place names

[Adam Schreiber]

On Mon, Mar 23, 2009 at 8:50 AM, Brian Quinion
<openstreetmap at brian.quinion.co.uk> wrote:
> On Mon, Mar 23, 2009 at 12:34 PM, David Earl <david at frankieandshadow.com> wrote:
>> On 23/03/2009 12:13, Adam Schreiber wrote:
>>> On Mon, Mar 23, 2009 at 8:08 AM, Elena of Valhalla
>>> <elena.valhalla at gmail.com> wrote:
>>>> finding a working algorithm that uses only objective data and is
>>>> always able to select the proper item to print would be excellent, but
>>>> also probably not feasible in real life.
>>> Population?  People can vote importance with their feet.
>> Population's not enough. Some places "punch above their weight" - the
>> example I like to use is Hay-on-Wye in Powys. It's under 2,000 people
>> but is most clearly an important market town for the surrounding area.
>> In population it would be a modest village, but it is more important
>> than that.
>
> I've been finding that the number of hotels (within the city area as
> defined by builtup_area file) is a surprisingly effective method of
> calculating the importance of a place, with the added advantage that
> it is data already present in OSM.

I like this idea.  For example, Clemson, SC's student population would
not be included in the US Census population data, but it arguably has
more hotels than Seneca, SC which is rendered above but has a larger
permanent population.  Given the large amount of government business
transacted in Albany, it probably has a lot of hotels too.

Cheers,

Adam