You are on page 1of 7

 Multidimensional data sources do not support aggregations and binned data.

 Tableau will allow you to save an invalid calculation; however, a red exclamation point appears
next to it in the Data pane
 Until you correct an invalid calculated field, you will not be able to drag it into the view.
 As you drag formula content to a shelf, card, or pane, Tableau provides visual feedback: A red X
indicates that you cannot drop the expression at the current location. If you drop an expression
that you have dragged from the calculation editor onto a field in the view, the expression
replaces that field. If the resulting expression is not valid, the resulting field will be red. You can
then edit the expression to make it a valid field.
 For purposes of calculations, the primary data source is whichever data source is currently
displayed in the Data pane.
 It is not possible to drag numeric bins, generated latitude and longitude fields, Measure Names,
or Measure Values into the calculation editor.
 You can copy and paste all custom defined fields, including calculated fields, ad hoc groups,
user filters, and sets between workbooks.
 Moving the calculation editor does not move Tableau Desktop, and moving Tableau Desktop
does not move the editor.
 When you are writing formulas, any part that displays in bold indicates that it will be computed
locally within Tableau on the aggregated results. Any normal weight text will be computed at
the database level.
 If you are connected to a Microsoft Analysis Services data source, the calculation editor
contains a "Run before SSAS" check box. Choose this option to execute the Tableau calculation
before any Microsoft Analysis Services calculations.
 Functions are colored light blue in formulas. Fields are colored orange in formulas. Operators
are colored black in the formula. Parameters are colored purple in formulas. Comments are
colored green in the formula.
 Auto-completion will not offer suggestions as you type a quoted string or if you begin by typing
a number. If you begin by typing an open square bracket, [, auto-completion shows fields,
parameters, sets, and bins, but not functions.
 Matches from secondary sources are only shown if there is an explicit blend relationship set
with the currently active sheet. Fields that are being used to blend the two data sources are
only shown once in the search results (the field shown is from the primary data source).
 Ad-hoc calculations, also known as type-in or in-line calculations, Ad-hoc calculations are
supported for fields on the Rows, Columns, Marks, and Measure Values shelves; they are not
supported for fields on the Filter or Pages shelves. Alternatively, you can double-click on an
empty shelf or on an empty part of a shelf to create a new ad-hoc calculation.
 Calculated members can either be new measures or new dimension members depending on
the hierarchy that you choose. For calculated dimension members, specify a parent member.
For calculated measure members, specify a result type. Finally select a solve order for the
member.( Sometimes a single cell in your data source can be defined by two different formulas.
The solve order defines the precedence given to each formula. Formulas with a lower solve
order are solved first. The default solve order is zero.)



  Ad-hoc calculations are also known as type-in or in-line calculations. Ad-hoc calculations are
supported on the Rows, Columns, Marks, and Measure Values shelves; they are not supported
on the Filters or Pages shelves. Typically you create ad-hoc calculations on-the-fly to do things
like,    Test a hunch,    Try a what-if scenario,    Debug a complex calculation. Ad-hoc calculations
are not available when you create groups, sets, hierarchies, or parameters. Ad-hoc calculations
are valid for creating trend lines, forecasts, and reference lines, bands, and distributions.
 When you are learning about table calculations, or experimenting with different options, a text
table is usually going to provide more intuitive insight than other chart types. In the example of
a running sum of product sales across several years, the addressing field is the Date field, and
the partitioning field is the product field.
 In the example of a running sum of product sales across several years, the addressing field is the
Date field, and the partitioning field is the product field. When you define the addressing for a
table calculation, all the other dimensions are used for partitioning.



 The dimensions that define the part of the table you are applying the calculation to (computing
along) are called addressing fields, and the dimensions that define how to group the calculation
are called partitioning fields. In the example of a running sum of product sales across several
years, the addressing field is the Date field, and the partitioning field is the product field. To
update addressing for a field in the view that already has a table calculation, right-click the field
and choosing one of the options under Compute using.
 Table(Across then down) will have only one partition. Tableau ignores nulls in rank calculations.
They appear as blank rows. Tableau ignores nulls in percentile calculations. They appear as blank
rows in a cross-tab and do not count towards the total number of items used in the calculation
(%).

 Level of detail expressions are computer in the data source. FIXED level of detail expressions
also ignore all the filters in the view other than context filters, data source filters, and extract
filters. In the case of a table-scoped level of detail expression, no scoping keyword is required.
 Level of detail expressions that reference floating-point measures can behave unreliably when
used in a view that requires comparison of the values in the expression.
 Level of detail expressions are not shown on the Data Source page.
 Sets cannot be used in the dimension declaration in a level of detail expression.
 With data blending, the linking field from the primary data source must be in the view before
you can use a level of detail expression from the secondary data source.
 FIXED level of detail expressions can result in measures or dimensions, INCLUDE and EXCLUDE
level of detail expressions are always measures.
 It is possible to define a level of detail expression at the table level without using any of the
scoping keywords.
 Level of detail expressions are always automatically wrapped in an aggregate when they are
added to a shelf in the view unless they’re used as dimensions.If you wrap a level of detail
expression in an aggregation when you create it, Tableau will use the aggregation you specified
rather than assigning one when any calculation including that expression is placed on a shelf.
When no aggregation is needed (because the expression’s level of detail is coarser than the
view’s), the aggregation you specified is still shown when the expression is on a shelf, but it is
ignored.
 http://onlinehelp.tableau.com/current/pro/online/windows/en-
us/help.htm#calculations_calculatedfields_lod_constraints.html%3FTocPath%3DAdvanced
%2520Analysis|Calculations|Level%2520of%2520Detail%2520Expressions|_____8


 By default, Tableau uses the entire table for percentage of. Percentage calculations are based on
the entire data source. The option you choose is applied uniformly to all measures that appear
on a worksheet. For instance, you cannot choose Percent of Column for one measure and
Percent of Row for another.
 If you place Measure Names as the inner dimension on the Columns shelf (that is, the dimension
farthest to the right), Tableau will return 100% for each mark because you cannot total up the
values for multiple measure names. For example, you can’t total up the values for SUM(Sales)
and SUM(Profit).
 If you place Measure Names as the inner dimension on the Rows shelf (that is, the dimension
farthest to the right on the shelf), Tableau will return 100% for each mark because you cannot
total up the values for multiple measure names. For example, you can’t total up the values for
SUM(Sales) and SUM(Profit).


 Tableau provides prediction bands which may be simulated or calculated from a closed form
equation. All models with a multiplicative component or with aggregated forecasts have
simulated bands, while all other models use the closed form equations. Trend is a tendency in
the data to increase or decrease over time. Seasonality is a repeating, predictable variation in
value.
 Therefore, if there is a six-month cycle in your monthly time series, Tableau will probably find a
12-month pattern that contains two similar sub-patterns. However, if there is a seven-month
cycle in your monthly time series, Tableau will probably find no cycle at all. Luckily, seven-month
cycles are uncommon.
 The original temporal method uses the natural season length of the temporal granularity (TG) of
the view. Temporal granularity means the finest unit of time expressed by the view. The new
non-temporal method, introduced with Tableau 9.3, uses periodic regression to check season
lengths from 2 to 60 for candidate lengths. The AIC of the five seasonal models and the three
non-seasonal models are compared and the lowest returned.
 When Tableau is using an integer dimension for forecasting, the second method is used. In this
case there is no temporal granularity (TG), so potential season lengths must be derived from the
data. The second method is also used if the temporal granularity is yearly. The second method is
also used for views with temporal granularity of minute or second.
 Ten seasonal models are estimated, five with a season length of 60 and another five with the
season length derived from the data. Whichever of the ten seasonal models or three non-
seasonal models has the lowest AIC, that model is used to compute the forecast.
 Since all selection is automatic when Tableau is deriving potential season lengths from the data,
the default Model Type of “Automatic” in the Forecast Options Dialog Model Type menu does
not change. Selecting “Automatic without seasonality” improves performance by eliminating all
season length searching and estimation of seasonal models.
 For Model Type “Automatic” in integer-, year-, minute- and second-ordered views, candidate
season lengths are always derived from the data whether or not they are used. Since model
estimation is much more time consuming than periodic regression, the performance impact
should be moderate.
 Multiplicative models can significantly improve forecast quality for data where the trend or
seasonality is affected by the level (magnitude) of the data
 Forecasting can be done on two date types, truncated dates and date parts, exact dates are not
valid for forecasting. In Date parts, one year part is required.
 Use the Ignore Last option in the Forecast Options dialog box to remove—or trim—partial
periods. The default is to trim one period.
 Tableau requires at least five data points in the time series to estimate a trend, and enough data
points for at least two seasons or one season plus five periods to estimate seasonality. For
example, at least nine data points are required to estimate a model with a four quarter seasonal
cycle (4 + 5), and at least 24 to estimate a model with a twelve month seasonal cycle (2 * 12).
 Tableau can only get more data when the aggregation for the measure you are forecasting is
SUM or COUNT.
 You can forecast quantitative time-series data using exponential smoothing models in Tableau
Desktop. With exponential smoothing, recent observations are given relatively more weight
than older observations. These models capture the evolving trend or seasonality of your data
and extrapolate them into the future.
 The date or time field can be discrete or continuous. In the Forecast Options dialog box, you can
specify how far into the future to estimate, the date part that the forecast is based on, and the
forecasting model to use.
 Forecasting requires a view that uses at least one date dimension and one measure. For
example:
 The field you want to forecast is on the Rows shelf and a continuous date field is on the Columns
shelf.
 The field you want to forecast is on the Columns shelf and a continuous date field is on the Rows
shelf.
 The field you want to forecast on either the Rows or Columns shelf, and discrete dates are on
either the Rows or Columns shelf. At least one of the included date levels must be Year.
 The field you want to forecast is on the Marks card, and a continuous date or discrete date set is
on Rows, Columns or Marks.
 You can also create a forecast when no date dimension is present if there is a dimension in the
view that has integer values.
 The lower the percentile you set for the confidence level, the narrower the prediction bands will
be.
 As with a date, when an integer dimension is selected to order the measures to be forecast, it is
no longer used to partition the data. When Tableau is using an integer dimension to forecast,
the Forecast Option and Forecast Description dialog boxes will automatically specify that
forecasting is aggregating by periods. Order in which the dimensions are considered for forecast,
columns, rows, pages and marks
 Forecast result type Trend—Show the forecast value with the seasonal component removed.
Quality—Show the quality of the forecast, on a scale of 0 (worst) to 100 (best). This metric is
scaled MASE, based on the MASE (Mean Absolute Scaled Error) of the forecast, which is the
ratio of forecast error to the errors of a naïve forecast which assumes that the value of the
current period will be the same as the value of the next period. The actual equation used for
quality is: 100* max ( 1-MASE,0). The Quality for a naïve forecast would be 0. The advantage of
the MASE metric over the more common MAPE is that MASE is defined for time series which
contain zero, whereas MAPE is not. In addition, MASE weights errors equally while MAPE
weights positive and/or extreme errors more heavily.
 How your prediction intervals are displayed depends on the mark type of your forecasted marks:
Line-Bands; Shape, Square, circle, bar or pie – Whiskers
 If there is more than one time series in your visualization, the Custom option forces them all to
be forecast using the same custom model. Constraining the models in this way usually results in
lower quality models than would be produced by automatic model selection.
 You cannot use a multiplicative model when the measure to be forecast has one or more values
that are less than or equal to zero, or even when some of the data points are too close to zero,
relative to other data points.
 You cannot specify a model with multiplicative trend and additive season because the result
may be numerically unstable.
 The information in the Describe Forecast dialog box is read-only, though you can click Copy to
Clipboard and then paste the screen contents into a writeable document.
 Because a forecast cannot be computed for a time series with null date values in the middle,
actions that filter data, either explicitly or implicitly can trigger a null forecast. The Keep Only
and Exclude commands on tooltips are examples of actions that can filter data implicitly—in
some cases, these commands are removed when a forecast is shown. Tableau also displays the
error message in the Forecast Options dialog box (Analysis > Forecast > Forecast Options).
Another indication is that the Forecast Indicator field on the Marks card shows Actual values,
but no Estimate values. http://onlinehelp.tableau.com/current/pro/online/windows/en-
us/help.htm#forecast_describe.html%3FTocPath%3DAdvanced%2520Analysis|Forecasting|
_____6
 Tableau can create a forecast when the forecast aggregation is a finer level of detail than the
view's aggregation, but not when it is at a coarser level of detail; even when it is finer, the two
values are only compatible if there is a strict hierarchy that Tableau can use (for example,
quarters can be evenly divided into three months, but months can't be evenly divided into
weeks)
 Forecasting is not possible when the result set from the query is too large. The limit is about
10,000 rows. All forecast algorithms are simple models of a real-world data generating process
(DGP). http://onlinehelp.tableau.com/current/pro/online/windows/en-
us/help.htm#forecast_how_it_works.html%3FTocPath%3DAdvanced%2520Analysis|
Forecasting|_____1


           
 Reference lines, bands, distributions, and boxes are not available when the view is a map using
online or offline maps. The higher the value you select for confidence interval, the wider the
bands will be. When you are displaying a line and a confidence interval, the shading will be
darker within the confidence interval, and lighter beyond it. When you are displaying a
confidence interval without a line, the fill colors are disregarded, though they are retained and
then applied if you decide later to show a line. Specify whether to Show recalculated line for
highlighted or selected data points.
 Distributions can be defined by confidence interval, percentages, percentiles, quantiles, or
standard deviation. In addition to the shading, you can add a line to mark a constant or
computed value along the axis.
 Quantiles - breaks the view into the specified number of tiles using shading and lines. When you
select this computation, you must also specify the number of tiles (from 3 to 10, inclusive).
Tableau uses estimation type 7 in the R standard to compute quantiles and percentiles.
 You can format the lines at the distribution boundaries (for example, style, thickness, and color)
as well as the fill gradient. Select from a list of predefined gradients. Select Symmetric to use a
single color instead of a gradient. You can also specify whether to add additional shading above
and below the defined distribution.
 Box plots provide quick insight into the distribution of data. Boxes indicate the middle 50
percent of the data (that is, the middle two quartiles of the data's distribution).
 You can configure lines, called whiskers, to display all points within 1.5 times the interquartile
range (in other words, all points within 1.5 times the width of the adjoining box. This is also
known as a schematic box plot.), or all points at the maximum extent of the data. This is also
known as a skeletal box plot.
 Specify whether to Hide underlying marks (except outliers)—that is, whether to hide all marks
except those beyond the whiskers.

      
 If you are using a multidimensional data source you can specify whether to compute the total on
the server using the underlying data or in locally using the data that you see in the table. The
default setting is to compute all totals on the server if you are connected to a Microsoft Analysis
Services data source, and locally if you are connected to an Essbase data source using the
aggregation specified in the cube. However, there are cases when it is not possible for the server
to compute the expected subtotals due to filtering or perfect pivoting.
 An additional value, Server, may be available. Server computation is not always available and
sometimes the totals will be blank for specific members in the view. When using server
computation keep in mind the following information. Server computation is only available for
ASO cubes. Server computation is not available for dynamic hierarchies. That means if the
members in the view are part of a dimension or hierarchy that is tagged as dynamic, they cannot
be included in the set of values you are using to compute the totals and will show up as blank in
the view. If you are computing totals for a calculated field whose formula makes assumptions
about other calculated members at different levels in the hierarchy, the totals will display as
blank in the view.
 Grand Totals cannot be applied to continuous dimensions.
 Two-pass total, the average you see in the grand total column is aggregated twice--once to
derive the column or row value, and then again across column or rows to derive the grand total.


 All - the parameter control is a simple type in field.


 Tableau confidence bands show upper and lower 95% confidence lines by default when you add
trend lines. Confidence lines are not supported for Exponential models.
 Force the y-intercept to zero option is available only when both the Rows shelf and the Columns
shelf contain a continuous field, as with a scatterplot.
 To add trend lines to a view, both axes must contain a field that can be interpreted as a number.
 Dates are interpreted as numbers for trendlines.
 For multidimensional data sources, the date hierarchies actually contain strings rather than
numbers. Therefore, trend lines are not allowed. Additionally, the ‘m/d/yy’ and ‘mmmm yyyy’
date formats on all data sources do not allow trend lines.
 However, trend lines cannot be turned on for stacked bars. You can turn off stacked marks by
clearing the Analysis > Stack Marks option.
 Trend line options are retained for the next time you turn on trend lines. However, if you close
the workbook with trend lines turned off, trend line options revert to defaults.
 Random errors are uncorrelated with each other and with the explanatory variable, and have
equal variance. Because a logarithm is not defined for number less than zero, any marks for
which the explanatory variable is negative are filtered before estimation of the model. The trend
line description reports how many marks were filtered before model estimation.
 With a polynomial model type, you also select a Degree between 2 and 8.more complicated
higher order polynomial models require more data to estimate.
 The second line, the R-Squared value, shows the ratio of variance in the data, as explained by
the model, to the total variance in the data.
 The third line, the P-value, reports the probability that the equation in the first line was a result
of random chance. The smaller the p-value, the more significant the model is. A p-value of 0.05
or less is often considered sufficient.
 The Assumptions required to compute (using Ordinary Least Squares) each individual trend line
are:
o     Your model is an accurate functional simplification of the true data generating process
(for example, no linear model for a log linear relationship).
o     Your errors average to zero and are uncorrelated with your independent variable (for
example, no error measuring the independent variable).
o     Your errors have constant variance and are not correlated with each other (for
example, no increase in error spread as your independent variable increases).
o     Explanatory variables are not exact linear functions of each other (perfect
multicollinearity).
 http://onlinehelp.tableau.com/current/pro/online/windows/en-
us/help.htm#trendlines_terms.html%3FTocPath%3DAdvanced%2520Analysis|Trend
%2520Lines|_____8
 If your data contains negative values Tableau cannot plot them on a logarithmic scale. All values
with a negative value will be displayed at 1 on the axis. You can then filter these records to
exclude them from the view.

You might also like