[postgis-users] Basic question #1 Managing temporality in spatial data - MAUP over time

[Ben Madin]

Thanks Paul,

I will check out contrib/seg - it sounds like it might have  
application for another (unrelated) issue I am dealing with. It's a  
neat solution (as are most of these modules) to a problem.

cheers

Ben


On 20/02/2009, at 5:01 AM, postgis-users- 
request at postgis.refractions.net wrote:

> Date: Wed, 18 Feb 2009 16:35:06 -0800
> From: Paul Ramsey <pramsey at opengeo.org>
> Subject: Re: [postgis-users] Basic question #1 Managing temporality in
> 	spatial	data - MAUP over time
>
> Your approach is fine.
>
> If you were dealing with much larger data collection, you'd find that
> over time your spatial selectivity starts to go down, as more and more
> data piles up in the same x/y area over time. That's when it's time to
> add an indexed time-range to your tuples, so that you can
> simultaneously index select on spatial and temporal ranges. For that I
> would suggest adding a time range column containing 'seg' objects (see
> ./contrib/seg) with the end values derived from unix timestamps.
>
> Paul
>
> On Wed, Feb 18, 2009 at 4:24 PM, Ben Madin <ben at remoteinformation.com.au 
> > wrote:
>> I hope no one minds this, but I have to deal with some changes to  
>> my data,
>> and thought it might be a good opportunity to ask the basic  
>> questions I've
>> been dwelling on for some time.
>>
>> Having just received a new set of the spatial data I am working with
>> (shapefiles of administrative regions), I am unsurprised to find  
>> that about
>> 15 districts have gone (been amalgamated), some new ones appeared,  
>> and
>> indeed one entire province removed!
>>
>> My intention was to add two columns to my gis data table -
>> geom_start_validity & geom_end_validity, and revise my queries (which
>> normally are analysing / displaying movement information) to choose  
>> the
>> appropriate geometry depending on the date range of the query.
>>
>> Is the approach sound? (Is this a reasonable idea? Am I completely  
>> off track
>> here? Will this make my table difficult to index? Should I find all  
>> the
>> polygons which are the same and delete the duplicates, extending the
>> validity dates?)
>>
>> (The interface which relates to the entry of movements would also  
>> be amended
>> to deliver a different set of provinces and districts depending on  
>> the date
>> of the movements. How I deal with movements that occur over the  
>> transition
>> is beyond me so far!)
>

-- 

Ben Madin
REMOTE INFORMATION

t : +61 8 9192 5455
f : +61 8 9192 5535
m : 0448 887 220
Broome   WA   6725

ben at remoteinformation.com.au



							Out here, it pays to know...