Database Reference
In-Depth Information
WARNING
In my experience, the connected scatterplot takes more effort up front to decipher what is
being shown, simply because we're less accustomed to seeing time shown in this way. If
we can't be present when the audience sees it, and if they're not that motivated regarding
the subject matter, then they'll be less likely to put in the effort and we should probably
stick with a chart type that will be easier to understand: one with time on the x-axis.
The baseball strikeout data was an easy “learner” data set for us because it was pre-aggreg-
ated and the only date type we needed to consider was year. In the real world, we often get
data in logs or long lists of records with a time-date stamp field. Let's consider how we can
explore changes over time with this type of data next.
The Date Field Type and Seasonality
Let's return to the New York City rat sightings data we considered in
Chapter 3
. Recall that
this data is formatted as a list of reported occurrences, and each row in the spreadsheet is a
single occurrence. The location, time, and date are reported, as well as a number of categor-
ical fields such as type of location. We considered the total number of rat sightings for each
borough, but we didn't consider how those sightings were spread out over time.
A number of questions could be asked of the list of records, such as:
▪ What time of the year do the most rat sightings occur?
▪ Is the number of rat sightings increasing, decreasing, or staying steady?
▪ How many rat sightings can we expect to see in the coming year?
Let's explore this data to find some answers. If you recall from
Figure 3-12
,
the rat sightings
data set included a field called
Created Date
, in the form of
mm/dd/yyyy
. Let's see how
Tableau treats a date field like this.
If we drag
Created Date
onto the Column shelf and then drag
Number of Records
from
the Measures area to the Rows shelf, Tableau creates a yearly line plot by default, as shown
in
Figure 9-14
.
