Time Series Analysis in Tableau
Chris Hui
VP of Product, Tracked
Data can be encoded via delimiters that can be split apart into their own fields (e.g. AUS-001)
Tableau has two distinct splitting options: Split or Custom split
Split: separates the data based off the most common delimiter
Data can be encoded via delimiters that can be split apart into their own fields (e.g. AUS-001)
Tableau has two distinct splitting options: Split or Custom split
Split: separates the data based off the most common delimiter
Cust Order Key | Split 1 | Split 2 | Split 3 |
---|---|---|---|
AA-B1-C1 | AA | B1 | C1 |
AA-B2-C2 | AA | B2 | C2 |
AA-B3-C3 | AA | B3 | C3 |
Data can be encoded via delimiters that can be split apart into their own fields (e.g. AUS-001)
Tableau has two distinct splitting options: Split or Custom split
Split: separates the data based off the most common delimiter
Cust Order Key | Split 1 | Split 2 | Split 3 |
---|---|---|---|
AA-B1-C1 | AA | B1 | C1 |
AA-B2-C2 | AA | B2 | C2 |
AA-B3-C3 | AA | B3 | C3 |
$$
Cust Order Key | Split 1 | Split 2 | Split 3 |
---|---|---|---|
AA-B1-C1.DD | AA | B1 | C1.DD |
AA-B2-C2.DD | AA | B2 | C2.DD |
AA-B3-C3.DD | AA | B3 | C3.DD |
SPLIT()
is helpful when you need to split up one or more delimiters in a textual field
The tokens (text) are returned either left to right (positive), or right to left (negative) dependent on position
Splitting functions are generally nested and combined with functions like:
IF()
/ TRIM()
/ CASE
By design, Tableau adheres to and recognizes ISO-8601 and Standard Gregorian date formats
However dates can sometimes appear in a textual / string format
The ISDATE()
function can be utilized to return True for all entries recognized as dates and False otherwise
DATEPARSE()
converts non-standard string fields to date time objects through explicit parsing of the date format
It can be utilized for all date strings that are recognized as strings as opposed to dates
MM-YY-DD might makes sense to us, but not to Tableau, unless you specify this with DATEPARSE()
ISDATE()
and DATEPARSE()
are useful to identify invalid dates, but not correcting these dates
DATETRUNC()
returns the lowest value for each date partition, dependent of the day specification (i.e. Week start on Monday)
DATETRUNC()
is useful for verifying re-occurring reporting periods for errors (e.g. Weekly)Time Series Analysis in Tableau