> On Nov 7, 2006, at 5:48 AM, [EMAIL PROTECTED] wrote:
>
> > From: [EMAIL PROTECTED]
> >
> > > Since that grows all the districts in parallel, but approximately
> > doubling districts at each step, at the end
> > > you'll have some districts with about double the population of
> > other districts.
> >
> >
> > Yea, I sorta said that at the end.
> >
> > I think that breaking the country up into simple blocks could allow
> > similar algorithms though.
>
> Such data is available.
> http://ftp2.census.gov/census_2000/datasets/Summary_File_1/
>
> For example, California is split up into 533163 little blocks. Only
> about 330,000 of them are actually populated. They cut the map into
> lots of little pieces.
I had a look at that too, and yeah, they are complex.
Block groups are a little less complex. However, the sometimes have a
block group contained entirely within another block group as an
island.
Another issue with blocks, was working out their geographic
information.
I managed to get both at the block group level, but not at the
block level.
Are blocks always completely contiguous ?
In any case, I think the real problem is that they collect alot of
data which is irrelevant to selecting districts. This is what
makes the file format so complex.
Also, are blocks actually hard to change ? If the are easy then
by modifying them, you can have some control of the end
districts. However, that would be a hell of alot less control
than being able to draw the lines any way a person wants.
What about just splitting the State into grid of 1km by 1km squares
and then allowing the legislature to (once off) move the boundaries
up to 500m in any direction. Also, they wouldn't be allowed
to cause a cross over of the boundaries or change the
connectivity. The end result is a grid that can for almost
all purposes be assumed to be a square grid. This makes
the algorithms easier to get to work, the data easier to understand
and a clear result from the algorithm. They wouldn't have to
deal with floating point geo data. The districting algorithm
would just assume a square grid and be given the population
per "square".
In practice, it would probably have to be defined in terms of
longitude and latitude rather than squares. For example,
the grid lines might be at 0.01 degree intervals running E-W
and N-S. If my calculations are correct, that gives around
1km edges.
However, some cities have a population density of greater
than 10k per square kilometer.
It might be worth allowing grid "squares" that have greater
than (100?) people to be further sub-divided. However, they
would be required to maintain proper connectivity.
It all comes down to how accurate the result needs to be. Also,
even if some boxes have a large population, other smaller
blocks can be used to even things out.
>
> > From: [EMAIL PROTECTED]
> >
> > > Since that grows all the districts in parallel, but approximately
> > doubling districts at each step, at the end
> > > you'll have some districts with about double the population of
> > other districts.
> >
> >
> > Yea, I sorta said that at the end.
> >
> > I think that breaking the country up into simple blocks could allow
> > similar algorithms though.
>
> Such data is available.
> http://ftp2.census.gov/census_2000/datasets/Summary_File_1/
>
> For example, California is split up into 533163 little blocks. Only
> about 330,000 of them are actually populated. They cut the map into
> lots of little pieces.
I had a look at that too, and yeah, they are complex.
Block groups are a little less complex. However, the sometimes have a
block group contained entirely within another block group as an
island.
Another issue with blocks, was working out their geographic
information.
I managed to get both at the block group level, but not at the
block level.
Are blocks always completely contiguous ?
In any case, I think the real problem is that they collect alot of
data which is irrelevant to selecting districts. This is what
makes the file format so complex.
Also, are blocks actually hard to change ? If the are easy then
by modifying them, you can have some control of the end
districts. However, that would be a hell of alot less control
than being able to draw the lines any way a person wants.
What about just splitting the State into grid of 1km by 1km squares
and then allowing the legislature to (once off) move the boundaries
up to 500m in any direction. Also, they wouldn't be allowed
to cause a cross over of the boundaries or change the
connectivity. The end result is a grid that can for almost
all purposes be assumed to be a square grid. This makes
the algorithms easier to get to work, the data easier to understand
and a clear result from the algorithm. They wouldn't have to
deal with floating point geo data. The districting algorithm
would just assume a square grid and be given the population
per "square".
In practice, it would probably have to be defined in terms of
longitude and latitude rather than squares. For example,
the grid lines might be at 0.01 degree intervals running E-W
and N-S. If my calculations are correct, that gives around
1km edges.
However, some cities have a population density of greater
than 10k per square kilometer.
It might be worth allowing grid "squares" that have greater
than (100?) people to be further sub-divided. However, they
would be required to maintain proper connectivity.
It all comes down to how accurate the result needs to be. Also,
even if some boxes have a large population, other smaller
blocks can be used to even things out.
Raphfrk
--------------------
Interesting site
"what if anyone could modify the laws"
www.wikocracy.com
--------------------
Interesting site
"what if anyone could modify the laws"
www.wikocracy.com
---- election-methods mailing list - see http://electorama.com/em for list info
