====Pivot Example: By Product Characteristic====

===Choosing Pivot Data===
Because pivoting makes new entries for every unique pivot value, we suggest that you only pivot on properties that **step rapidly** between values, or you take advantage of functions such as [[math_round|Round]] to remove decimal places.

The best pivots are non-sensor data, such as **batch number**, **product code** or other inputs from your production scheduling systems.

===This Example===

In this example, we're going to use a [[module_pivot|pivot analytic]] to see how quickly our production line is creating different //thicknesses// of product.

This will let us know if thickness is relevant, and which particular thicknesses are causing problems.

===Available Information===

^Asset^Property^Desc^
|Paint Line|Status - Painting|Has a value of **1** when painting, and **0** when stopped.|
|Paint Line|Speed - Actual|The speed of our product line|
|Paint Line|Thickness|The thickness of the product being processed|

===Initial Layout===

{{start_pivot.png}}

To begin with, we have one [[pivot_pivot|Pivot]] node named //Pivot On//, and one [[pivot_output|Pivot Output]] node named //MinMaxAvg//

These are both fed by [[ardi_point|ARDI Point]] nodes, bringing data in from ARDI asset properties.

===Step 1: Choose the Pivot===

Our first step is to choose //what we are going to pivot on//.

  * Click on the //ARDI Point// node that feeds the //Pivot On// node.
  * Choose **Paint Line Thickness** from the list
  * Right click the //Pivot On// node and choose 'Title' from the list
  * Enter the name 'Thickness' as the name of the pivot.

{{pivot_analogue_pivot.png}}

===Step 2: Add the Speed Measurement===

  * Click on the //ARDI Point// already connected to the //MinMaxAvg// node.
  * Choose **Paint Line Speed - Actual** as the property.
  * Right-click the //MinMaxAvg// node and choose **Title**.
  * Enter the name 'Speed'

{{pivot_analogue_complete.png}}

===Step 3: Ignoring Transitions===

If we pressed **Save and Visualise** right now, we'd have a working analytic. But there would be some issues.

Although the //thickness// value from our system changes abruptly between the old and new value, the process itself takes time to adjust itself to the new product. This is called a **transition**.

This gradual movement from the old product to the new causes our 'max' and 'min' values for each product to bleed into one-another slightly. Although the //average// values are good, but the maximum and minimum values also include the speeds from the product before and the product after.

{{catching_transition_edges.png}}

Luckily, we can see the transition is happening when the lines **status** goes from 'painting' (1) to 'stopped' (0).

Pivot analytics also include an [[pivot_ignore|Ignore Node]]. If the input to this node is ever //true// or //1//, any values will simply be ignored and not assigned to any pivot value.

==Adding The Ignore Node==

  * Right-click in the empty space on the graph, and choose **Add Node**, **Report** and **Ignore**.

==Add the Status==

Next, you'll need to add the status. 

Although you can use a normal [[ardi_point|ARDI Point]] node and a comparison, we've included an easy way to deal with digital values with the [[ardi_status|ARDI Status]] node, which has dedicated outputs for **true** and **false**.

   * Right-click in the empty space on the graph and choose **Add Node**, **ARDI** and **ARDI Status**.
   * Left-click where it says //None// on the new node, and choose the **Paint Line Status - Painting** property from the list.
   * Connect the second output (named //false//) on the new node to the input of the Ignore node.

{{pivot_analogue_ignored.png}}

===Analytic Complete===

We now have a complete analytic that shows us the different speeds based on product thickness, ignoring those disruptive product transitions.

{{pivot_analogue_display.png}}

===See More===

If there wasn't any pre-made method of detecting a transition, you might want to instead [[pivot_timers|use a timer to ignore a window of time]].