Method for making tables and charts that embed statistically significant difference indicators inside their cells/descriptions and inside column headers as a legend
-
Figure out what the (standardized) structure of the data would need to look like in order for us to then pipe it into visualization packages (and develop example code that demonstrates how to use it) we might need to just combine the level text with the relevant indicators, separated by
\nto put them on a new line - the formatting of this could be reconfigured later on, like to not put it on a new line and/or to enclose in parentheses- aside from embedding the indicators, if only diffing two periods, could embed the percentage point difference from one period (that you’d set) to the currently displayed one (like showing 2023 results generally, with significant differences from 2022)
- also formatting labels in a time series chart differently based on the labeled value’s significant differences from one of the other included values. Different modes of formatting for these states: no sig diff, any sig diff, sig higher, sig lower. They should likely be able to stack, too, to indicate any sig diff first (like putting a border around the labels) and then which direction the difference is in (green for higher, red for lower).
- similarly, just formatting the values in a table based on sig diff direction (blue for higher, red for lower)
- output tables to HTML so they can be copied and pasted into reports, slide decks, Excel workbooks as needed
-
Get example output photos
- make them
- put screenshots of them into here