Standardize colors in legends

[eriliawu]

Standardize colors for sub-group analysis. i.e. make sure colors in the by-group analysis are consistent across different analysts’ project. (we can talk about this in our Monday morning meeting)

[EmilHafeez]

https://cran.r-project.org/web/packages/viridis/vignettes/intro-to-viridis.html ?

[EmilHafeez]

please see template code change below.

Replace

scale_color_manual(name="Open Time \nAfter Labeling", labels=c("12","15","18","21","24","27","30"), values=c("hotpink","olivedrab3","red","orange","grey","purple","#13B0E4")) +

with

scale_color_viridis(discrete=TRUE, option="turbo",labels = c("Lunch (11:00-13:59)", "Afternoon (14:00-16:59)", "Dinner (17:00-20:59)", "Evening (21:00-23:59)")) + guides(col=guide_legend("turbo"), size=guide_legend("turbo")) +

[eriliawu]

• For meal time
  • keep lunch, afternoon, dinner and evening in the figures
  • keep all 6 meal time categories in the table reporting
• macro nutrients
  • ordering: carb, protein, fat, sugar, sat fat, fiber, sodium
• item category:
  • burrito, dessert, other entree, side, taco (beverage comes last if this category is included)

[eriliawu]

• covariate balance
  • CBPS, IPTW, unmatched

[kritisingh21]

this code (provided above in sample) -> guides(col=guide_legend("turbo"), size=guide_legend("turbo")) -> results in the legend title being "turbo" so I changed it to "Food Category" for both instances which "fixes" (for a lack of a better word) the legend title....not sure if this correct, but please let me know!

[EmilHafeez]

Solution below:

First, Create a named object where each category of the interesting variable is assigned a Hex color value from the viridis color palette. The only subanalysis with differing number of factors is the by-item-category, so this only applies there. All other analyses should use the viridis color palette but do not need manual tuning. Since Lloyd's analysis includes beverages, he has six total factors, and other analysts will manually set the number of factors in the below code to replicate this, thereby keeping colors consistent.

food_category_name_vector = c("Burrito", "Dessert", "Other Entree", "Side", "Taco", "Beverage")

ggplot_palette_named = setNames(object = scales::viridis_pal()(length(unique(food_category_name_vector))), nm = unique(food_category_name_vector))

Then, use the object in the scale_color_manual values option into the ggplot script such that each named category gets the assigned Hex colors.

scale_color_manual(name = "Food Categories", labels=c("Label 1", "Label 2"), values = ggplot_palette_named) +

Note that this requires we have the factors in a consistent order between analysts, as previously discussed 6/29/22 above.

[EmilHafeez]

color_reference= as.data.frame(c(2,3,4,6,8)) colnames(color_reference) <- c("color_categories") ggplot_palette_named = setNames(object = scales::viridis_pal()(length(unique(color_reference$color_categories))), nm = unique(color_reference$color_categories)) ggplot_palette_named=ggplot_palette_named[-6]

[EmilHafeez]

Please note that the adjustment to retrieve the colors from turbo rather than base viridis is as follows

ggplot_palette_named = setNames(object = scales::viridis_pal(option = "turbo")(length(unique(color_reference$color_categories))), nm = unique(color_reference$color_categories))