Everywhere in ThoughtSpot where you can Auto Analyze with ThoughtSpot, you can also configure a customized SpotIQ analysis. The customized analysis can be a one-time customization or can apply to all future analysis of that particular data source.
Finding customization controls
Every menu or location where you can Auto Analyze with ThoughtSpot, you can also choose Custom Analyze. It looks like this:
Regardless of where you are in the ThoughtSpot UI, the SpotIQ Customize Analysis dialog has the same layout. The columns, of course, are specific to the data you launched the dialog from.
This dialog has three tabs, Included columns, Excluded columns , and Customize Algorithms. The Customize Algorithms is by far the most advanced panel.
The SpotIQ insights page has an option for this dialog and also an option to customize a single analysis.
The Improve Analysis dialog always excludes a value from all future analysis.
Finally, you can also set some analysis options from your user Profile.
You’ll notice only one of these options is actually related to analysis. The others relate to how you receive notifications about SpotIQ operations.
Eliminate null values
Null values should be eliminated from your analysis.
From next analysis | From future analyses |
---|---|
In the Search bar, add a filter |
Set Exclude null values from my analysis on your user Profile to exclude them from any future SpotIQ analyses you do. |
Exclude uninteresting column(s) in insight
You should always exclude columns you are not interested from your analysis.
From next analysis | From future analyses |
---|---|
Exclude a column using Improve Analysis on a visualization. Or set the value through the Excluded columns tab of the Customize Analysis panel. |
Choose Also exclude from all my future analysis on the Excluded columns tab of the Customize Analysis panel. Alternatively, exclude using the Improve Analysis dialog on an insight that uses the column you want to exclude. |
Include an interesting column
You can always includes columns that interest you in your analysis.
From next analysis | From future analyses |
---|---|
Choose Customize Analysis and select to Include columns that you want to appear. |
Ensure **Index Priority** is between 8-10 on the column through data modeling. |
Remove known date outliers
Your data may contain known outliers. For example, you are in the middle of a quarter and only want to analyize the previous quarter. Anything from the present quarter could contain an outlier.
From next analysis | From future analyses |
---|---|
In the Search bar, add a filter |
Not applicable. |
Too few insights
Your SpotIQ analysis may not provide you as many insights as you think it should.
From next analysis | From future analyses |
---|---|
Choose Customize Analysis and Customize algorithms Decrease the Multiplier for Outlier Detection to a value closer to zero. |
Not applicable. |