Visualizing
Rank
and Size of Cities and Towns
Part II: Greater London, 1901-2001
Sandra
Arlinghaus
and Michael
Batty
Dr. Sandra Arlinghaus is Adjunct Professor at The University
of
Michigan, Director of IMaGe, and Executive Member, Community
Systems
Foundation.
Dr. Michael Batty is Bartlett Professor of Planning at
University
College London where he directs the Centre of Advanced Spatial
Analysis.
Please
set screen to highest
resolution and use a high speed internet
connection.
Please download the most recent free
version of Google
Earth®.
Make sure the
"Terrain"
box in Google Earth® is checked.
Download
the following file to use in Google Earth®:
|
Greater
London: A
Century of Change
Greater London is
composed of the City of London (of quite small population)
and 32
boroughs that surround the central city.* As in Part I, we begin looking at changes
in the data
sets of interest, by decade, over the course of the
20th century. Rank-size plots are shown in Figure 1;
the general
pattern is as one might expect. There appears to
be a change in pattern around the time of World War II.

Figure
1. Rank-size plots of the City of London
and 32 surrounding
boroughs composing Greater London. Click here to view a
.mov file
in which the reader can control the animation rate.
|
To take a closer look
we separate the rank-plots into two sets, in Figure 2.
Figure 2a
shows the plots from 1901 to 1941 and Figure 2b shows them
from 1951 to
2001.

Figure
2a.
|

Figure
2b.
|
Parallel to the Part I case, we note that any given locale
is
likely to change rank over time. Thus, we look at the
data set in
relation to 1901 ranks, for the entire century (Figure 3a)
and for the
pre-and post-World War II data (Figures 3b and 3c).
The general
pattern appears quite wild while the shorter time span ones
centered on
either side of World War II offer a more organized
picture. Is
that picture more organized for Greater London than
it is for the entire UK? These observations are
perhaps not
surprising. They do benchmark strategy and might offer
interesting visualizations to those doing policy, planning,
or
historical studies of the study region.

Figure
3a.
|

Figure
3b.
|

Figure
3c.
|
Next, we map the data.
The
Google Earth® screenshots
of Figure
4 show not only
all the population bars for each borough and for the City of
London for
1901 but also for each of the other decades up through
2001.
Again, we have animated them so the reader can quickly see
such
change. Click on any single image in Figure 4 (a-k) to
see a
larger image. Or, keep track of up to nine changing
scenes on the
screen at a single time. To drive around, download the
associated file
used to make the images
(Figure 4l).
There are a number of
interesting patterns one can observe; we invite the reader
to add to
these or to challenge them.
- Boroughs close to the central city are larger earlier
and larger
as a group in pre-World War II Greater London. The
general
pattern is pyramidal with the apex close to the City of
London.
- Post -World War II sees a flattening of the heights of
parallelepipeds across the entire region.
- The last two decades begin to see some growth back
toward the
center.
- Early on, the southeast boroughs seemed under-sized in
relation
to other bars; later, that changes.
It might
be
interesting to compare and contrast this situation for
London with
other major cities, both in the UK and elsewhere, especially
in regard
to movement patterns in relation to war. Indeed, one
might
consider applications for
this method for other urban areas in order to study land use
planning,
circulation, and infrastructure in relation to
disasters.
Tower
Hamlets:
A Local View.
The borough
of Tower
Hamlets is adjacent to the City of London: it is a
"close-in"
borough. Simple animation of the rank-size graph
easily shows its
changing population/size and rank pattern over time (Figure
5).
In addition, animation from Google Earth® makes it
easy to
compare and
contrast the relative rise and fall in population of Tower
Hamlets in
relation to Barnet, a "far out" borough (Figures 6a and 6b;
again, to
take a closer look at either model, click on the image to
link to a
bigger file). Thus, scholars investigating patterns
associated
with sprawl might find this tool to be helpful in a variety
of ways.

Figure
5.
|
A visual
limitation in perspective is involved with this
procedure. One
cannot
see changes over time while driving around within the virtual
distribution of a single time slice. The animation
scheme is
useful
because it is hard to retain 3D models in the mind and
mentally
superimpose one time frame on top of another. The
strategy
developed above, while apparently useful in many ways, does
not allow
one to see simultaneously the full picture and also see change
over time. There may be other strategies that fulfill
that
need.
Future Directions
Both authors have recently offered a number of different
strategies for
visualizing data sets over time and also from different
periods of
time. In addition, one might imagine that a host of
other
possibilities will arise given the relative ease of current
remarkable
visualization techniques.
Add
Sound
In order to merge the spatial and temporal concerns, we
consider first
introducing audio files to supplement the visual. Click
on any of
the boroughs in the map below. A sequence of notes from
a musical
scale will play. They represent the rise or fall in rank
of that
borough during the twentieth century. Different boroughs
will
play different notes from the musical vectors serving as a
basis for a
musical vector space in which both rank and size change
through
time. As the
reader listens to change over time he/she is free to study
simultaneously spatial aspects of the map Generally, the
pattern
of
the notes works as follows:
- a musical
vector
that is relatively high in pitch throughout is one whose
associated
region has had relatively high rank throughout the time
period (and
vice versa).
- within
a
musical vector, be it generally high, low, or middle, the
higher notes
represent higher numerals (hence lower ranks) and vice
versa.
The method
of
construction of the musical vectors, including much detail,
appears in
Appendix II. Click on the musical map of Figure 7 and
listen to
the rise and fall of rank...a guide that those who have
vision disabilities may be able to employ.

Figure
7. Musical map of Greater London. Click on a
borough or
the City of London and listen to the general rank
pattern and to rise
and fall of rank within that general pattern.
|
Change
the Geometry
The methods
for
looking at spatial change over time outlined above in the
context of UK
data sets offer exciting prospects for imaginative geometric
use of the
internet. What they all have in common is that they are
couched
in Euclidean geometry. The most radical, and perhaps the
most
interesting, approach might well be to change the geometry--to
employ
the non-Euclidean. In the last issue of Solstice, we announced
our
interest in this topic and outlined a research agenda for
using
non-Euclidean geometry to look simultaneously at spatially
disparate
rank-size plots from different locales, time frames, or
both. To
that agenda it now seems important to add that we should
investigate
the role of internet mapping and geometry, especially as they
draw from
Google Earth®. Might
one
imagine the Google Earth® "sphere" as a
rotating Poincaré Disk on which to embed
non-Euclidean views of rank-size plots? Stay tuned...the
answers will be coming soon!
APPENDIX
PROCEDURE USED WITH "A MUSICAL
GENERATOR®"--DOWNLOAD A
FREE DEMONSTRATION COPY AND OPEN IT TO FOLLOW ALONG WITH THE
DISCUSSION
BELOW.
- Create a matrix showing change in rank, over time, of a
city or a
set of cities. We choose "Greenwich" for the sake of
example of
procedure. The row associated with Greenwich will be
referred to
as its "vector."

|
- Enter data from the row associated with Greenwich into the
generator.
- Direct approach:
- Choose the tab named "Data"
- Click on the letter "N" in order to directly enter
numerals
associated with Greenwich.
- Type the numerals, leaving a space between successive
entries, creating a space-delimited file.
- Click "OK" when done. You will then see a small
chart
appear in the previously blank left area of the
window. The chart
will have the label "data." Change the title to
"Greenwich" by
right-clicking and choosing "rename."
- Indirect approach: bring in data directly from
Microsoft
Excel®
(or other software) using directions from the help files
of A
Musical Generator®.
- Next, generate music from the data.
- Click to highlight channel 8; it has longer-sounding
notes
associated with it than does channel 1.
- Drag the chart entitled "Greenwich" and drop it on top
of the
graphic on the "Notes" button.
- Then, hit the "play" button to hear the raw sound of
audio
associated with the data for Greenwich.
- Adjust the music. We give the setttings used in
the files
for the clickable map of Greater London.
- Set the Tempo to 182: slide the bar.
- Set the number of measures to 10; there are 11 entries
in the
vector.
- Click on the "Duration" button and set the "Maximum"
to 33
(the number of possible ranks) and the "Default" also to
33.
- Click on the "Notes" button. Set the "Minimum"
and
"Maximum" to correspond with the minimum and maximum
values of the
numerals in the rank vector for Greenwich. A
Musical Generator®
allows values from "c" as the Minimum to "g10" as the
Maximum. We
use the following assignment pattern to associate
musical note value
with rank value, from 1 to 33, assuming after
considerable
experimentation that a musical octave, based on Western
style with a
"Major" tone scale, is presumed to begin with "c".
- c3=33; d3=32; e3=31; f3=30; g3=29; a3=28; b3=27;
c4=26;
d4=25; e4=24; f4=23; g4=22; a4=21; b4=20; c5=19; d5=18;
e5=17; f5=16;
g5=15; a5=14; b5=13; c6=12; d6=11; e6=10; f6=9; g6=8;
a6=7; b6=6; c7=5;
d7=4; e7=3; f7=2; g7=1. Thus, to cover the entire
range of ranks,
one would set the Minimum in the "Edit notes aspect"
window to c3 and
the Maximum to g7--as an absolute maximum and absolute
minimum for the
rank situation.
- To focus on the general nature of the Greenwich
vector,
however, we restrict the focus to the local maximum and
local minimum
of the vector itself. The minimum is 13 and the
maximum is
20. Thus, set the Minimum in the "Edit notes
aspect" window to b5
(assigned to 13) and the Maximum to b4 (assigned to
20). Now, try
playing the associated music once again.
- Save your work both as "Greenwich.tmg" and as
"Greenwich.mid"--the latter is a midi file which plays
on the internet
and elsewhere.
RELATED REFERENCES
See links on author names in title material for links to
publication
lists.
- A Musical Generator 3.0,
from
MuSoft
Builders, http://www.musoft-builders.com/
Last
accessed Nov. 27, 2006.
- Arlinghaus, Sandra and
Batty,
Michael.
2006. .Zipf's
Hyperboloid? Research
Announcement, Solstice:
An
Electronic Journal of Geography and Mathematics,
Volume
XVII, No. 1.
- Arlinghaus, Sandra and
Arlinghaus, William. 2005 Spatial
Synthesis (Chapter 2, scroll to end for music
characterizing
central place hierarchies). Ann Arbor, MI: Institute
of Mathematical Geography.
- Batty,
Michael. 2006: Rank clocks, Nature, Vol. 444, 30 November,
2006,
doi:10.1038. Link to
reprint.
*The City of London population data from
1901 to 1991 is
City of
London 26882
19619 14158
11054
5324 4767
4245
5864
5900
4000
Before 1901 the population was likely much higher; indeed, in
1801 the City of London probably had the largest population in
the
United Kingdom. London lost more than
half its population in the interwar years. By 1951 the
population
was very low, never
to recover, as it was all employment by then.
Solstice:
An Electronic Journal of Geography and Mathematics,
Volume XVII,
Number 2
Institute of Mathematical Geography (IMaGe).
All rights reserved worldwide, by IMaGe and by the authors.
Please contact an appropriate party concerning citation of
this
article:
sarhaus@umich.edu
http://www.imagenet.org