|
|
|
|
Home
/ Com Products / Product Guide / Choosing a Graph
Bar graph
|
|
Bar graphs compare values to each other.
This is an example of a simple bar graph. It shows a single
data set, one data point per bar. |
|
|
|
A bar graph consists of two or more parallel bars of equal width drawn on an
X-Y grid. Bar graphs compare amounts to each other. They can also suggest
trends, especially in vertical form.
-
2D bar graphs are available in five styles: simple, stacked, stacked
percentage, stacked floating, and pareto. 3D bar graphs can be drawn in all of
those styles plus an additional z-clustered style. See the example above and
those on the next page.
-
You can draw bars either vertically or horizontally. In the vertical format,
viewers tend to attribute a left-to-right sequence to the bars, whether you
intend one or not.
-
By default, Graphics Server places bars at increments of 1, starting at 1. In
rare cases, you may want to specify the positions of bars along the X axis (for
vertical bars) or Y axis (for horizontal bars).
|
|
 |
When the simple style is used to chart multiple data sets, data
points within a set are clustered together along the X axis.
The first point in a set is drawn on the left of the group.
Subsequent points are drawn to the right of the preceding ones.
|
|
The Z-clustered style groups data points within a set along
the Z axis. The first point in a set is drawn in the back rank.
Subsequent points are drawn in front of the preceding ones.
With this form, a tall bar can easily obscure shorter bars behind
it. The graph may be tilted or rotated to bring all data points
in view.
|
|
|
|
With the stacked style, an entire data set forms a single bar.
Data points are represented as segments of the bar. The first
point is the base of the bar. Subsequent points are stacked
on top of preceding ones. Negative data points are not shown.
|
|
The stacked percentage style graphs each data point as a percentage
of the sum of all values in its set. This style highlights the
relative weight of a point within its set. It is best used when
comparisons between sums of data sets are irrelevant or undesirable.
|
|
|
|
Floating style is like stacked except that the first data point
in each set is invisible, causing the bar to appear to float
above the X axis. Negative data points are not shown.
|
|
Pareto style graphs a single data set. Bars are sorted in order
of descending value, regardless of the order in which data is
entered. This style is often used in quality control applications
because it highlights primary contributors to a problem so that
they can be addressed first.
|
Use a bar graph...
-
In simple form when you want to compare values within a set.
-
In simple clustered form when you have multiple sets and want to highlight the
progression of values within a set.
-
In z-clustered form when you want to highlight the progression of values
between sets.
-
In stacked form when you want to compare the sums of multiple sets.
-
In stacked percentage form when comparing actual values is less important than
comparing relative weights of values within a set.
-
Drawn vertically when you want to imply a trend to the values shown on the
bars.
-
Drawn horizontally when you don't want to suggest a trend.
Other graph types to consider
-
If you're graphing only one data set and you want to compare parts to the whole
rather than to each other, use a pie graph.
-
When trends are your primary concern, and especially when you want to imply a
gradual flow between data points, consider a line
graph or area graph.
-
When you have many sets of data and a large number of points within each set,
consider a surface graph.
|
|
|
|