Monday, 8 April 2013

Horse arrives after cart! More Census stuff

I have used the title above to reflect my current view that I should have done this post first.  It is a common outcome of the "let's suck it and see" approach to analysis used to generate my first post in this series that a number of interesting issues arise which require redoing step one.

That makes my self esteem head for the basement so perhaps I will follow a fiction-writing model and call this a prequel!

What is a region?
This is a very basic point since it determines the geographic extent about which I am going to write.  In Regional Science.  There are two broad types of Region which could be used:

  • Nodal region, comprising a central point of service provision and the areas which depend on it. They are possibly easier to define in a quantified way because one can find out to where the market area for a given node extends.  However they sometimes:
    • overlap.  For the area I am going to be talking about most people would regard Queanbeyan as their node for retail shopping, but Bungendore as the node for waste disposal. (Sorry Bungendore, but the tip is about the only attraction you have for me.  Oh: and the Mudchooks); and
    • are hierarchic.  This is illustrated by education where 
      • a primary school may be in a village or locality;
      • High schools tend to be in towns; while 
      • Universities are in cities.

  • Homogenous region, an aggregation of areas which share some common attribute. The areas are typically contiguous but I don't see this as necessary.  The attribute could be based upon: 
    • public administration (eg an LGA);
    • physical features (eg watershed boundaries);
    • cultural attributes (eg the catchment area of a newspaper).
My principle interest in this stroll down the avenue of analysis is the homogenous region defined by the catchment of the Stoney Creek Gazette.  This is mainly because that is 'my patch': i live there and know a lot of the residents.  I shall return to this below.

Other homogenous regions that could be of interest could include

  • the Palerang Council LGA - which includes most but not all of Carwoola as well as the other two components of the Gazette area;
  • The Molonglo Valley - which includes much of the three areas I am reporting on but excludes (I think) the Wanna Wanna area (which drains to the Queanbeyan River). 

The main study area.
When I did the first post I had basically opened up Table Builder and started off looking at Carwoola.  This is partly what I mean by the 'suck it and see (SIAS - one needs a few obscure acronyms) approach".  It wasn't pure SIAS because I did have a question in mind- "what proportion of the residents of Carwoola were retired from the workforce?"

You may wonder why I didn't read the very helpful Table Builder User manual.  That is sort of like asking me why I don't bother with aps when driving somewhere.  However I have hung out with IT guys and realise that when all else fails one reads the manual - or at least consults it for the solution to whatever problem the initial failure to read it has caused.

In this case I was wonder if it was possible to add data for areas together.  This revealed the wonders of Custom Geography!(If you are interested they are explained in section 4.5 of the manual.).  So I created a Custom area comprising the three State Suburbs of Carwoola, Hoskinstown and Primrose Valley.  Part of the resulting screen is shown below.

I just happened to see the tab labelled "Map View".   So I whacked that and got ...
Yowza!  What a great improvement.

So my next step was to get some data.

Age Profiles
I had previously produced some age profiles for the three areas separately.
With the exception of the high proportion of 10-19 year olds in Hoskinstown the profiles are quite similar.  However I can now do a profile for the whole area.
That still shows the bulge around 40 - 60 years so I decided to produce an age-sex profile to see if the sex bias was still evident for the 20 - 30 year olds.
Yes, it is still there, which is probably to be expected as Carwoola contributes 70% of the population for the combined area.

No comments: