To fix inconsistencies in analytic settings, functionality needs to be enhanced to ensure they are on par with other visual component enhancements made for Cartesians, and additional settings for sc...
I don't even understand why this doesn't work this way already. It makes sense that the value the trend line is mapping would need to be numerical, but what does it matter how it's split up? If I have 5 number values in a chart, you shouldn't need to know what I call those groups in order to calculate a trend line... it's already divided into 5 discrete datapoints which you can use to generate a trend line. If your calculation needs numbers to map against, just use the column number! For example, if I have a chart that that shows Page Views for: "March"(55) "April"(65) "Cucumber"(42) "Pickle"(99) You don't need to calculate the value of "Pickle" to map a trend line, just read those values in as (1,55) (2,65) (3,42) (4,99) and calculate the trend line!
