Add Trend Lines
Trend lines are lines added to visualizations to highlight and predict patterns across multiple values in a set of data. For example, you may want to track your company's overall pace of revenue growth over the past 5 years.
This article shows how to add, edit, and delete trend lines in visualizations.
Summary of Content
Requirements
Trend Line Model Types
Linear
Logarithmic
Exponential
Power
Polynomial
Quadratic
Determine Best Fit from R² Values
Add a Trend Line
Edit a Trend Line
Delete a Trend Line
Axis Scale Types
Related Resources
Requirements
 To edit a visualization, you must have Can Edit or Can Explore access to the individual workbook.
 The visualization must have values plotted on both its XAXIS and YAXIS. Both axes must also have a compatible scale type applied. The following scale types are compatible with trend lines: Linear, Time, Log, Pow, Sqrt.

Trellised visualizations, stacked visualizations, and visualizations with values plotted on the COLOR field do not support trend lines.
Trend Line Model Types
The following model types are available: Linear, Logarithmic, Exponential, Power, Quadratic, and Polynomial.
If you’re unsure which model type to choose, we recommend trying a few and picking the one that looks best with your data. A line’s R² value can also help you determine best fit.
Each model type is based on an underlying linear regression formula. These formulas are included below. However, understanding the math behind the line is not necessary for using trend lines in Sigma.
LinearLinear trend lines are used for data that follows a simple, steady, straight line. Data points may increase or decrease, but the trend remains steady. Formula: Y = a + b * X 

LogarithmicLogarithmic trend lines are used for data sets in which the rate of change increases or decreases quickly before leveling out. Formula: Y = a + b * log(X) 

ExponentialExponential trend lines are used for data sets in which the values rise or fall at constantly increasing rates. Exponential trend lines don't recognize zeros or negative numbers. Formula: Y = a * e^(b * X) 

PowerPower trend lines are used for data sets in which the values increase at a specific rate. Power trend lines don't recognize zeros or negative numbers. Formula: Y = a * X^b 

PolynomialPolynomial trend lines are used when data fluctuates. They're helpful when analyzing gains and losses over a large period of time. Sigma's polynomial trend lines default to a polynomial order of 3 and support orders 3  7. Polynomial order refers to the number of coefficients applied. This effects how many hills or valleys are present in the line. The higher the number, the more hills or valleys to expect. Order 3 trends typically have one or two hills or valleys. Whereas order 2 trends would have only one hill or valley. Formula: Y = a + b * X + … + k * X^3 

QuadraticQuadratic trend lines are 2ndorder polynomial trend lines. Much like Polynomial trend lines, they're used to smooth out fluctuations in a data set. Formula: Y = a + b * X + c * X^2 
Determine Best Fit from R² Values
Rsquared (R²) represents how well the trend line fits the data. This is based on variance between data points.
R² values are always between 0 and 1. Values closer to 0 signal that the line fit is worse, while values closer to 1 signify a better fit.
To include the R² value on a visualization's trend line, check the trend line's Show value option in the editor panel. The value will appear wherever the trend line label is positioned in relation to the line.
Add a Trend Line
To add a trend line to a visualization:

Select the visualization.
To support trend lines, a visualization must have values plotted on both its XAXIS and YAXIS. Both axes must also have a compatible scale type applied. The following scale types are compatible with trend lines: Linear, Time, Log, Pow, Sqrt.
Visualizations that meet any of the following do not support trend lines: trellised visualizations, stacked visualizations, and visualizations with values plotted on the COLOR field.
The column you want to use to calculate a trend should be plotted on the YAXIS.  Open the visualization's format panel.
 Click TREND LINES.

Click + Add new.
 Click the new Trend line dropdown.

Under Select column, select a column to use for the trend.
The available list contains only columns plotted on the visualization's YAXIS.

Under Model, select a trend line model type:
 Linear: display as a bestfit straight line
 Logarithmic: used for data sets in which the rate of change increases or decreases quickly before leveling out
 Exponential: used for data sets in which the values rise or fall at constantly increasing rates
 Power: used for data sets in which the values increase at a specific rate.
 Polynomial: used when data fluctuates. They're helpful when analyzing gains and losses over a large period of time. Sigma's polynomial trend lines default to a polynomial order of 3. This value is configurable below.

Quadratic: 2ndorder polynomial trend lines used to smooth out fluctuations in a data set
Polynomial order effects how many hills or valleys are present in the line. The higher the number, the more hills or valleys to expect.
For no label, leave the text box blank or uncheck Label text.
R² represents how well the trend line fits the data based on variance between data points.
Edit a Trend Line
To Edit an existing trend line, enter Edit mode, then:
 Open the visualization's format panel.
 Click TREND LINES.

Select the dropdown associated with the trend line.

Make any of the following edits:
 Linear: display as a bestfit straight line
 Logarithmic: used for data sets in which the rate of change increases or decreases quickly before leveling out
 Exponential: used for data sets in which the values rise or fall at constantly increasing rates
 Power: used for data sets in which the values increase at a specific rate.
 Polynomial: used when data fluctuates. They're helpful when analyzing gains and losses over a large period of time. Sigma's polynomial trend lines default to a polynomial order of 3. This value is configurable below.
 Quadratic: 2ndorder polynomial trend lines used to smooth out fluctuations in a data set

To select a new column to use for the trend, pick a column under Select column.
The available list contains only columns plotted on the visualization's YAXIS. 
To change the line's model type, select a model type under Model.
 [optional for Polynomial trend lines only] Under Degree, select a polynomial order between 3 and 7.
 Select the line style: Solid, Dashed or Dotted.
 Select a new line size or color.

To add a text label check the Label text checkbox, and type into the text box that appears below.
For no label, leave the text box blank or uncheck Label text.

To position or reposition the label in relation to the line, select a position from the dropdown menu: Top right, Top left, Bottom right, Bottom left.

To show/hide the line's R² value, check/uncheck Show value.
R² represents how well the trend line fits the data based on variance between data points.
Delete a Trend Line
To delete a trend line, enter Edit mode, then:
 Open the visualization's format panel.
 Click TREND LINES.
 Select the dropdown associated with the trend line.

In the bottom right corner of the trend line editor, click Delete.
Axis Scale Types
A visualization only supports trend lines if its columns are plotted on both its axes and if a trend line compatible scale type is applied. The following scale types are compatible with trend lines: Linear, Time, Log, Pow, Sqrt.
Visualizations that use an incompatible scale type display a warning message in the TREND LINES section of the format panel. To clear this warning and add a trend line, check your axes’ scale types and update as appropriate.
To change an axis' scale type:
 Open the visualization's format panel.
 Open the axis' format section: XAXIS or YAXIS.

Select an option from the Scale type dropdown menu.
Not all visualizations support all scale types. The following scale types are trend line compatible:
 Linear: Plots data along the axis using a linear numeric scale
 Time: Plots data along the axis as time values
 Log: Plots data along the axis using a logarithmic scale
 Pow: Plots data along the axis using a power scale

Sqrt: Plots data along the axis using a squareroot scale
Related Resources
Intro to Building Visualizations