Database Reference
In-Depth Information
walk you through producing a simple visualization that shows how easy it
can be to create graphs that extract meaning from your data.
Tableau is a drag-and-drop environment; if you don't like using the mouse,
you might just have to put away your mouse-o-phobia for a little bit if you
want to create fantastic pictures. In order to graph anything, you need at
least two fields; the graph lets you see how the first field varies in relation to
the second one. This is true for almost any kind of graph: A bar graph plots
a name against a height, and a line graph plots an ordered series against
another value. Even a pie chart needs at least two variables: the names of the
pie slices and their sizes.
You're actually not limited to plotting two variables in Tableau; you can use
color, size, or even tooltip text as other dimensions. And you're not limited
to pie, line, or bar graphs either; if you have geographic data, you can easily
plot the data on a map in dozens of different ways. You're also not limited
to values that are actually present in raw form in your data. Tableau can
perform a number of different types of aggregations:
SUM()
,
COUNT()
, and
so on and can apply them automatically to make the graph make more
sense. Probably the best way to figure out what is available is just to play
with it for a while and see what you come up with.
As a demonstration, however, we can create a bar chart of the word count
of each Shakespeare corpus in the Shakespeare dataset. The fields from the
table are divided up into dimensions and measures; the dimensions are the
text fields and the measures are the numeric fields. Tableau automatically
categorizes fields into dimensions and measures based on statistics about
the fields such as number of unique values and whether they are numeric
types. For the Shakespeare table, the dimensions are
corpus
and
word
;
the measures are
corpus_date
and
word_count
. There are a couple of
other special dimensions and measures, such as Measure names, Number of
Records, and Measure Values, which aren't in the actual dataset but can be
computed on-the-fly. You can drag the measures and dimensions between
the boxes to create visualizations.
Did Shakespeare Get Lazy as He Got Older?
As a way of demonstrating the use of Tableau on BigQuery data, let's see if
we can figure out whether Shakespeare got wordier as his career progressed,
or whether he got lazy and started wrapping up his plays as soon as possible.
You could, of course, do this with a query; but because the results are
