Improvements to the interaction chapter

[malcolmbarrett]

• simulation for ATE vs ATT/ATC
• after simulation, triple d'agostino plot (Add a figure for examining treatment heterogeneity (residuals vs fits split by treatment) #147)
• add plots e.g. https://marginaleffects.com/chapters/interactions.html as well as show how to get data to manually plot
• Add example of calculating effect modification
• Add example of calculating causal interaction

[malcolmbarrett]

@LucyMcGowan we are missing examples of both effect modification and causal interaction! We only show the functional form version.

Reading the chapter gave me an idea, though: maybe calendar quarter for effect modification (it's something we definitely can't control) and ticket season for interaction (it's correlated with calendar date but Disney CAN change it and we could imagine that there really is a joint effect of EMM + Value or something... but I think maybe there was some sparsity across some of the groups so we have to check I guess)