Upon first entering the development screen, one of the first items most users notice is the different icons used to denote different data objects, and the colours: blue and green.
You could be forgiven for assuming that blue = dimensions and green = measures, after all, this is how they normally appear.
However, this is only part-right:
Blue = Discrete data
Green = Continuous data
But, what is continuous and discrete?
Discrete data is data that persists throughout the set, such that the object can be used to categorise your measures
Whereas Continuous data is that which flows from 0; it may not necessarily be sequential, but it does continue
A great example of the two types is that of Date:
Dates make for excellent examples for the very reason that they can be equally difficult to consider:
Discrete
Example
Year
2000
Month
May
Day
02
Weekday
Monday
Time
15:31:43
Week Number
18
This is because all but Year occur each year period for example (yyyy-mm-dd):
Look at these two dates: 2002-02-18 & 2003-02-18
Both could be categorised to:
18th February
February
Day 18 of February
Day 49 of the year
Contrast this to continuous data:
2002-02-18: The complete date is continuous, as we enter the 19th (2002-02-19), we are now no longer able to categorise data from a future period against that of the previous period.
So, why is this important?
First-off, it will help you to understand the visual cue that Tableau Software have built into Tableau but second and most importantly, it will help you to understand how Tableau is working and why you cannot plot multiple lines against side-by-side bar charts:
Two discrete dimensions on the columns
2. One discrete dimension
3. One discrete dimension
4. One discrete dimension with date as Continuous
When more than one definition object (dimension) has been added, Tableau has no choice but to plot the data-items as discrete, this too is the same for the line which you would ideally prefer to plot over time: Tableau is simply plotting by Month
Upon removal of the Date dimension, Tableau plots the values aggregated to only the Category object
With the removal of the category from the layout, this is what chart 1 looks like but with the addition of the Category to break-down the bar value
Converting the Date dimension to continuous has allowed a continuous plot over time but broken-down by category; removal of Category from the columns will result in the same chart as Chart 3, but plotted over a discrete period
Icon Explanation
Data Pane
Real-time data-source connection: same symbol used for all data-types eg: database, flat-file, Excel, clipboard
Extract: Data is a snapshot from a different source. Your workbook size will include the extracted data and, you must refresh this source as regularly as the original source to remain present
Published data-source: Connection could be either real-time or extracted but runs exclusively from a Tableau Server or Tableau Online
Folder: Used for object organisation only
Geo: Data has been identified as having a global function eg country, state, town etc
String data.
Hierarchy: Objects have been organised into an expanding / contracting hierarchy
Continuous Date
Continuous Date-Time
Numeric
Bin: Continuous data have been categorised for counts: needed for histogram
Grouped: One or more discrete measures have been grouped to create a new measure
Set: User-defined filter has been created and retained for future use
Boolean
Continuous Date
Continuous Date-Time
Equals sign (demonstrated with a numeric symbol here for clarity) indicates the object is a calculation
Analytics Pane
Summarise - Functionally equivalent to reference lines
Constant Line: Add a Tableau generated constant line to the chart - you can modify the a constant value
Average Line: Adds an average line to either cell, pane or table
Median with Quartiles: Adds median and quartiles reference lines, both of which can be modified
Box Plot: Use to generate box and whiskers reference lines
Add the grand totals to the viz
Model - Functionally equivalent to reference lines
Average with 95% CI (Confidence Interval)
Median with 95% CI
Trend Line: Plot a configurable trend line through your data
Add a configurable forecast
Cluster: Add clustering from a dimension or groups of dimensions
Custom - Functionally equivalent to reference lines
Open the reference lines dialogue to the lines tab
Open the reference lines dialogue to the bands tab
Open the reference lines dialogue to the distributions tab
Open the reference lines dialogue to the box and whiskers tab