Graphics Reference
In-Depth Information
1.1. Data: Analog and Discrete Data
ous (representing real values, for example, in the
interval [0, 5]). The numbers called ordinals
represent the rank order (1st, 2nd, 3rd etc.) of the
data. Examples are data such as economic status
of some people (low, medium and high), a mineral
hardness scale, or the horse race results.
In terms of a scale of variables, the spacing
between the values may not always be the same
across the levels of the data. When the intervals
between the values of data are equally spaced and
the differences between pairs of measurements
can be compared, the variable is called an interval
variable. When we can determine a lowest value
of the data (such as an absolute zero) we can ap-
ply mathematical operations of addition, subtrac-
tion, multiplication, division, and exponentiation
(Ward, Grinstein, & Keim, 2010; UCLA Academic
Technology Services, 2012; Wikiversity, 2012).
Coordinate systems (with the x and y axes)
should contain clear labeling and a scale with
regular intervals that tells about the amount of
change (Tufte, 1983, p. 14). Graphs must not
quote data out of context. Without axes and scales,
a graph does not answer “compared to what?”
and does not tell if the change is big, what hap-
pened before and after our measurements, is it a
seasonal, temporary, or serious change, and how
this change compares with events in other places
(Tufte, 1983, p.74-75).
Time-series charts are the most frequently used
form of graphic design for complex statistical
material, e.g., thousands numbers about weather.
Time series charts can explain big data sets where
changes in recorded values go along to the regular
rhythm of seconds, minutes, hours, days, weeks,
months, years, centuries, or millennia. About 75%
of all graphics published were the time-series.
The problem with time-series is that the pas-
sage of time is not a good explanatory variable:
descriptive chronology is not causal explanation
(Tufte, 1992).
See Table 1 for Your Visual Response.
Data in the form of incoming signals such as light,
temperature, electrical, mechanical, pneumatic,
hydraulic or other signals can be conveyed in
an analog or digital format. An analog signal
changes continuously over time; we can measure
small fluctuations in its quantity. For example, a
microphone diaphragm responds to sound (which
means changes in air pressure) and causes changes
in a voltage or a current in an electric circuit.
Analog signals have higher than digital resolution
because digital resolution goes in discrete, separate
steps. A digital system uses discrete changes in
information, so signals are conveyed as electronic
or optical pulses with their amplitude presented
as a logical 1 (pulse present) or a logical 0 (pulse
absent). Digital audio or digital photography use
such binary numeric system.
Data can be described as categorical, ordinal,
or arbitrary. A categorical variable of the data
(also called a nominal variable) can take on non-
numeric values with two or more categories, for
example colors. The ordinal, ranked values have
an implied ordering (for instance, a mild, medium,
or hot salsa). The arbitrary variables can take on
any range of values without ordering (for example,
geographical location in a country represented
by that country's international telephone access
code, or the make or model of a car). One cannot
clearly arrange the data, for example by adding or
subtracting them, as well as we cannot say, 'less
than' or 'greater than' about the nominal num-
bers. For example, gender, marital status, race,
religious or political affiliation, college major,
and birthplace are examples of categorical data
measured at a nominal level.
In ordinal measurement the ordered, numeri-
cal data are called the ordinal or rank data. They
can take on numeric values that are binary (with
values of 0 and 1 only), discrete (taking on integer
values, sometimes being in a subset), or continu-
Search WWH ::




Custom Search