Remove/replace nicePlot feature

[fontikar]

Need to consider whether this is still useful. @dfalster suggests not. But would be good to offer a plotting alternative.

[fontikar]

@dwarton agrees with @dfalster on this one. He adds that users can plot the sma.obj however they like with whatever package they want. Shall we also close this one?

[dfalster]

Hi @fontikar, If we're removing the feature, we can't just close, first need to remove the code. Then revaluate the plot.sma function. I think we should provide some guidance on how to easily plot, using both baseR and ggplot. Can you look into it see more and then discuss?

[fontikar]

Ah yes apologies @dfalster I wasn't clear in my previous comment. I didn't mean close the issue as away to remove the feature. From what I gathered @dwarton suggested to just leave the feature alone as it works as a quick visualisation and matches figures presented in the original MEE paper. I agree perhaps a tutorial on plotting with baseR or ggplot2 would be valuable! I will investigate these and will come back with some options!

[dfalster]

Got it. In that case

• review other all functions should be exported. Perhaps fewer than current
• develop ggplot equivalent, as existing code is already base R

[fontikar]

Sounds great! I think we can definitely reduce number of exported functions for sure!

[dwarton]

did I suggest leaving it? I can't remember... I think plot of an sma object should do something, maybe that is what I meant.

[fontikar]

Have a very simple version of the function going, can produce the smatr plots for single S(MA), next is to expand to multiple groups 317d0fc My plan is to use group_by compute the line and the plot but grabbing group info from obj$coef in a function is proven tricky, maybe a for loop but I think map is possible!

[dfalster]

Good start Fonti.

For multiple groups, The output of Sma already includes a table of pars, I think it's called groupsummary (?)?

Happy to review together.

[fontikar]

Managing well to expand to multiple groups with common slope, next steps to get add the ability to varying elevation and slope sma(y~x*group)

[fontikar]

Great, current functions work well for:

• sma(y~x*group) use cases
• sma(y ~ x+group, type = "elevation")
• sma(y ~ x+group, type = "shift")

Next steps:

• Setting a default for sma(longev ~ lma, slope.test = 0.8, data=leaflife), obj$group == "all"
• Use cases when (S)MAs are forced through the origin
• Residual plots
• qq plots