Geography 215: QUANTITATIVE METHODS
Dr. Rodrigue
Graded Lab 3: SPATIAL MEASURES OF CENTRALITY AND DISPERSION
LAB EXERCISE A: Mean center, standard distance, and Manhattan median
All of the questions in Lab 3A have to do with the point data distribution in
Figure 1. The relevant reading in McGrew and Monroe is Ch. 4.
Figure 1: Distribution of C. foetidus, Sulphur Mt., Ventura Co.,
CA
quadrat grid in meters
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
___________________________________________________
| * | 0
| * * | 1
| * | 2
| * | 3
| * * * | 4
|* * * * | 5
| | 6
| * | 7
| ** | 8
^ | | 9
| | |10
N | |11
| | * * * ** |12
|* * |13
| |14
| |15
| |16
| |17
| * |18
| * * |19
| * * |20
| * |21
| |22
| |23
| * * |24
-------------------------------------------------- 25
- Using your spreadsheet, calculate the x-y coördinates of the mean
center of the plants mapped in the quadrat. Mark the mean center on the map
in a contrasting color or point design.
- Mean x coördinate: _______________ in m from upper left (NW)
corner
- Mean y coördinate: _______________ in m from upper left (NW)
corner
- Calculate the standard distance for this point pattern. On the map, draw
a circle with that radius on the map around the mean center.
- Sd = _______________ m
- What can normally be found within that circle?
___________________________________
- _________________________________________________________________
- What actually turned up in this distribution?
_____________________________________
- _________________________________________________________________
- Calculate the relative distance for this distribution. Helpful hints:
Area of a circle = pi x squared radius. Area divided by pi gets you the
squared radius. Take its square root and, voilà, the necessary radius
for doing relative distance.
- Rd = _______________
- The relative distance is nearly the analogue of the co-efficient of
variation, but in two-dimensional space. That is, the co-efficient of
variation is to the standard deviation as, roughly, relative distance is to
standard distance. The analogy is not exact, however. The co-efficient of
variation is calculated by dividing the standard deviation by the mean. Why
isn't relative distance calculated simply by dividing the standard distance by
the mean center?
- _________________________________________________________________
- _________________________________________________________________
- _________________________________________________________________
- _________________________________________________________________
- What are some shortcomings of dividing the standard distance by the radius
of a circle containing the same area as the study area? In other words, why
do you have to be a little bit cautious about interpreting differences in
relative distances among distributions measured in different areas?
- _________________________________________________________________
- _________________________________________________________________
- _________________________________________________________________
- _________________________________________________________________
-
Calculate the Manhattan median for the same points, using the orientation of
the implicit grid provided (N/S - E/W). Mark the Manhattan median directly on
Figure 1.
- In your own words, discuss the difference between the general concepts of
Euclidean distance and Manhattan distance. Which one pretty consistently
produces a larger value than the other? Under which circumstances would the
Euclidean distance and the Manhattan distance between two points be equal?
- _________________________________________________________________
- _________________________________________________________________
- _________________________________________________________________
- _________________________________________________________________
LAB EXERCISE B: Manhattan median and Euclidean median
The following questions have to do with Figure 2, which depicts the Census
tract boundaries within the small railroad town of Escalpón, Nevada.
The town is a nearly homogeneous middle income community of nearly 40,000
souls. Each of the six Census tracts has about 6,500 to 6,750 people (so
there's no real point to weighting the tracts for this exercise) and encloses
one square mile.
Figure 2: Escalpón, NV
^
| to Tonopah
---------
|
|
--------- |
| A |
^ | ---------|---------
| | | B |
N | | |
| |--------- | |
| C | |
| ---------|---------
| D |
| |
--------- |
|
|---------| ---------
one mi.
| to Las Vegas
v
- Calculate the Euclidean distance between the centers of each of the six
Census tracts and each of the four locations labeled A through D. This will,
then, entail 24 separate measurements. Please express your measurements in
miles, using the scale provided, and round your calculations to two decimal
places.
A to 1: __________mi. B to 1: __________mi.
A to 2: __________mi. B to 2: __________mi.
A to 3: __________mi. B to 3: __________mi.
A to 4: __________mi. B to 4: __________mi.
A to 5: __________mi. B to 5: __________mi.
A to 6: __________mi. B to 6: __________mi.
.............................
C to 1: __________mi. D to 1: __________mi.
C to 2: __________mi. D to 2: __________mi.
C to 3: __________mi. D to 3: __________mi.
C to 4: __________mi. D to 4: __________mi.
C to 5: __________mi. D to 5: __________mi.
C to 6: __________mi. D to 6: __________mi.
- Which of the four locations minimizes the aggregate Euclidean distance
between itself and the cores of the six Census tracts? That is, which of the
four choices is closest to the unknown Euclidean median of this distribution
of Census tract centers?
- _______________
- Now, assume that all roads in Escalpón run parallel or
perpendicular to the railroad shown, that is, they form a perfect,
Manhattan-style grid. Show the Manhattan median for the four labeled
locations directly on Figure 2.
last revised: 09/11/98
© Dr. Christine M.
Rodrigue