Most types of Insights that can be created in Insight have the ability to perform custom binning or transformation on the variables used in the chart or graph. These settings are accessed through the Variable Settings button that appears on the first tab of most Insight definition dialogs.
Within the Variable Settings button opens the Variable Settings dialog. In general, the dialog provides the ability to define settings for two variables involved in the Insight: the Analysis Variable and the Series Variable (note that some insights may not involve both of these variables). For each variable, options are presented for binning and transforming the variable prior to it being used in creating the chart output. The options available depend on the type of variable:
- Categorical variables, like the example of Gender on the left of the screenshot below
- Numeric variables, like the example of Age on the right of the screenshot below.
Each set of options is described below.
Options for Categorical/String Variables
- Max Categories to Show – This determines the maximum number of categories of the variable to show in the chart. This does not include NULL or 'others' categories as defined below in other options. The maximum value that can be entered is 25. If the variable has more categories than the value you enter, some of the categories will be classified as "Other" (see below).
- Include NULL as a Category – check this box to allow null values to be included as an additional category.
- Create “Others” Category – Checking this will allow any extra categories (i.e. not accounted for in Max Categories to Show above) to be grouped into an additional category labeled “Others”.
- "Others" Definition - this defines which categories are labeled as "Others".
- Minimum Frequency - the categories with the smallest count of records will get grouped as "Others".
- Minimum Metric Value - the categories with the smallest value of the summary metric will get grouped as "Others".
- Ordering – Select the desired method of ordering the categories on the axis.
Options for Numeric Variables
- Bucketing (Binning) Method – select the method of grouping for the analysis variable.
- None - this setting can be useful for integer data with relatively small number of values. If the variable has too many values for display, binning will be applied.
- Equal Sized - this setting will bin the variable in a way that creates the same number of records in each bin (note that the length of each bin will be different).
- Equal Length - this setting will bin the variable in a way that creates bins that are of the same length, likely with a different number of records in each bin.
- Custom Cutpoints - this setting lets you manually enter cutpoints for binning in the "Custom Cutpoints" box below.
- No. Buckets – select the number of buckets/bins to be created in the case of Equal Sized or Equal Length settings.
- Custom Cutpoints – If the Custom Cutpoints method was selected above, enter the cutpoints you wish to use, separated by commas. The cutpoints you enter will be included based on a ≤ sign. Note that values below and above the extreme endpoints in your list will be included as well. Therefore, if you enter 6 values in the list, the result will be 7 bins created.
- Include NULL as a Category – check this box to allow NULL entries to be grouped into an additional category.
- Remove Outliers – when using the Equal Length bucketing method, check this box to remove outliers to ensure that extreme values do not affect the binning.
Click Save to save your settings and exit the dialog. Click Cancel to cancel any changes you have made to the settings.