109
How Is Non-Spatial Data Handled by GIS?
about the spatial properties of the objects in that layer, but what it does not
have is any other information about those objects (such as the tax-assessed
value of the house, the house’s address, the name of the owner, the number
of people living in the house, the material the driveway’s made of, or how old
the driveway is). All of these attributes represent other non-spatial information associated with each of the objects. When representing attribute data in
GIS, these values can take one of four forms: nominal, ordinal, interval, or
ratio data.
Nominal data are values that represent some sort of unique identifier.
Your Social Security number or telephone number would both be examples
of nominal data—both of these values are unique to you. Names or descriptive information that are associated as a location would be nominal data, as
they’re unique values, the same as a value from a classification scheme. Also,
the difference between numerical nominal values is not significant—you
can’t add your Social Security number to a friend’s number and come up with
a relative’s number—the same way you can’t subtract your phone number
from one friend’s number and come up with another friend’s phone number.
Ordinal data is used to represent a ranking system of data. If you have
data that is placed in a hierarchy where one item is first, another is second, and
another is third, that data is considered ordinal. For instance, if you had a map
of the city and were going to identify the locations of the homes of the winners
of a local car-racing event, the points on the map would be tagged as to the
location of the first-place winner, the second-place winner, etc. Ordinal data
deals solely with the rankings themselves, not with the numbers associated
with these ranks. For instance, the only values mapped for the car-race winners is their placement in the race, not their winning times or the cars’ speed,
or any other values. The only thing ordinal data represents is a clear value of
what item is first, which is second, and so on. No measurements can be made
of how much better the first-place winner’s time was than the second-place
winner, only that one driver was first and one driver was second.
Interval data is used when the difference between numbers is significant, but there is no fixed zero point. With interval data, the value of zero is
just another number used on the scale. For instance, temperature measured in
degrees Celsius would be considered interval data since values can fall below
zero, and the value of zero only represents the freezing point of water, not the
bottom of the Celsius temperature scale. However, since there is no fixed zero
point, we can make differences between values (for instance, if it was 15 degrees yesterday and 30 degrees today, we can say it was 15 degrees warmer),
but dividing numbers wouldn’t work (since a temperature of 30 degrees is not
twice as warm as a temperature of 15 degrees).
Ratio data are values with a fixed and non-arbitrary zero point. For instance, a person’s age or weight would be considered ratio data since a person
cannot be less than zero years in age or weigh less than zero pounds. With
ratio data, the values can be meaningfully divided and subtracted. If we want
to know how much time separated the car-racing winners, we could subtract
the winning driver’s time from the second-place driver’s time and get the
attributes the nonspatial data that can
be associated with a
spatial location.
nominal data a type
of data that is a unique
identifier of some
kind. If numerical, the
differences between
numbers are not
significant.
ordinal data a type of
data that refers solely
to a ranking of some
kind.
interval data a type of
numerical data in which
the difference between
numbers is significant,
but there is no fixed
non-arbitrary zero point
associated with the
data.
ratio data a type of
numerical data in which
the difference between
numbers is significant,
but there is a fixed
non-arbitrary zero point
associated with the
data.
How Is Non-Spatial Data Handled by GIS?
about the spatial properties of the objects in that layer, but what it does not
have is any other information about those objects (such as the tax-assessed
value of the house, the house’s address, the name of the owner, the number
of people living in the house, the material the driveway’s made of, or how old
the driveway is). All of these attributes represent other non-spatial information associated with each of the objects. When representing attribute data in
GIS, these values can take one of four forms: nominal, ordinal, interval, or
ratio data.
Nominal data are values that represent some sort of unique identifier.
Your Social Security number or telephone number would both be examples
of nominal data—both of these values are unique to you. Names or descriptive information that are associated as a location would be nominal data, as
they’re unique values, the same as a value from a classification scheme. Also,
the difference between numerical nominal values is not significant—you
can’t add your Social Security number to a friend’s number and come up with
a relative’s number—the same way you can’t subtract your phone number
from one friend’s number and come up with another friend’s phone number.
Ordinal data is used to represent a ranking system of data. If you have
data that is placed in a hierarchy where one item is first, another is second, and
another is third, that data is considered ordinal. For instance, if you had a map
of the city and were going to identify the locations of the homes of the winners
of a local car-racing event, the points on the map would be tagged as to the
location of the first-place winner, the second-place winner, etc. Ordinal data
deals solely with the rankings themselves, not with the numbers associated
with these ranks. For instance, the only values mapped for the car-race winners is their placement in the race, not their winning times or the cars’ speed,
or any other values. The only thing ordinal data represents is a clear value of
what item is first, which is second, and so on. No measurements can be made
of how much better the first-place winner’s time was than the second-place
winner, only that one driver was first and one driver was second.
Interval data is used when the difference between numbers is significant, but there is no fixed zero point. With interval data, the value of zero is
just another number used on the scale. For instance, temperature measured in
degrees Celsius would be considered interval data since values can fall below
zero, and the value of zero only represents the freezing point of water, not the
bottom of the Celsius temperature scale. However, since there is no fixed zero
point, we can make differences between values (for instance, if it was 15 degrees yesterday and 30 degrees today, we can say it was 15 degrees warmer),
but dividing numbers wouldn’t work (since a temperature of 30 degrees is not
twice as warm as a temperature of 15 degrees).
Ratio data are values with a fixed and non-arbitrary zero point. For instance, a person’s age or weight would be considered ratio data since a person
cannot be less than zero years in age or weigh less than zero pounds. With
ratio data, the values can be meaningfully divided and subtracted. If we want
to know how much time separated the car-racing winners, we could subtract
the winning driver’s time from the second-place driver’s time and get the
attributes the nonspatial data that can
be associated with a
spatial location.
nominal data a type
of data that is a unique
identifier of some
kind. If numerical, the
differences between
numbers are not
significant.
ordinal data a type of
data that refers solely
to a ranking of some
kind.
interval data a type of
numerical data in which
the difference between
numbers is significant,
but there is no fixed
non-arbitrary zero point
associated with the
data.
ratio data a type of
numerical data in which
the difference between
numbers is significant,
but there is a fixed
non-arbitrary zero point
associated with the
data.