Information Technology Reference
In-Depth Information
The three types of scales that appear in graphs based on dimensions as we know
them in Yahoo! Web analytics are as follows:
Nominal Scales
nominal scales are lists of items that belong to one shared dimension
but do not relate to each other in any way other than that. Therefore, the order of this
list has no meaning. This could for example be dimensions such as Country, City, and
browser Type.
Ordinal Scales
ordinal scales are lists of items that belong to the same dimension. and
opposite from what we know from nominal scales, ordinal scales do have a unique
order. This could, for example, be dimensions such as Weeks, Day of Week, and Page
Views per Visits.
Interval Scales
interval scales are lists of items that, like ordinal scales, belong to the
same dimension and have a unique order, but they represent a value as well. an inter-
val scale is a quantitative scale, such as the sum of visits in a month, which is then
divided into a set of smaller ranges, perhaps a day, thus showing the sum of visits per
every day in that month.
244
For nominal and ordinal scales, i recommend that you go with the bar chart and
only use trend charts for interval scales. There is little lost and no major harm done in
using bar charts on all three scales, but from a visualization point of view, it is wrong
to use trend charts on nominal and ordinal scales.
adding a trend chart involves filling in the usual suspects, but with the notable
difference of a top group (see Figure 9.31).
Figure 9.31
Adding a trend chart dashboard item
We wanted to generate a trended revenue line graph in Figure 9.31, and added
in a top group based on the dimension search engines. remember that you can easily
have multiple visualizations based on the same data on the dashboard. For example,
you could have revenue This Month as part of a KPi box and then still have a trended
revenue line graph to the right of it.
The result of the values added in Figure 9.31 appears in Figure 9.32.
