We need a way of defining values that should be treated as blank or null. Since when you spray a blank field as numeric it is stored as 0 and internally they will assign certain values as null/ignore/blank values.
So for plugins that calculate values (univairiate, percentiles, etc) we need a way to build an ignore list for each column.
So we have a few options. Either on each plug-in where it makes sense we need to have a UI added to build a list of ignore values. (we would also want a way to to copy the same ignore values to all fields)
Option two a separate plugin that functions like global variables and can be left disconnected. In this plugin you would define the ignore values here. These would then need to be referenced in the calculations.
Option three add a third tab to the dataset plugin that adds the ignore list per column to that plugin.
I'm open to any of the three but tend to lean to the separate plugin. In any case the functionality will need to handle the following cases.
List of discrete values to ignore
basic boolean logic (ignore if >95) ignore if < 0)
Ignore if starts with specific character ( myCol[1] = '.')
This filtering would only effect the values on that column so if you have two columns x,y
x,y
3,4
0,3
-1,3
and you wanted to filter x<0 then the average calculation would be
We need a way of defining values that should be treated as blank or null. Since when you spray a blank field as numeric it is stored as 0 and internally they will assign certain values as null/ignore/blank values.
So for plugins that calculate values (univairiate, percentiles, etc) we need a way to build an ignore list for each column.
So we have a few options. Either on each plug-in where it makes sense we need to have a UI added to build a list of ignore values. (we would also want a way to to copy the same ignore values to all fields)
Option two a separate plugin that functions like global variables and can be left disconnected. In this plugin you would define the ignore values here. These would then need to be referenced in the calculations.
Option three add a third tab to the dataset plugin that adds the ignore list per column to that plugin.
I'm open to any of the three but tend to lean to the separate plugin. In any case the functionality will need to handle the following cases.
List of discrete values to ignore basic boolean logic (ignore if >95) ignore if < 0) Ignore if starts with specific character ( myCol[1] = '.')
This filtering would only effect the values on that column so if you have two columns x,y
x,y 3,4 0,3 -1,3
and you wanted to filter x<0 then the average calculation would be
x ----- 3+0/2=1.5 y ----- 4+3+3/3 = 8