plot_trends

This function can be used to visualize the results of the rank_change function, or to otherwise compare values from two different model versions or sets of estimates. This can help streamline the vetting process. All characteristics for each estimate should be matched, e.g. year, location, age. plot_trends creates plots of the estimates over some unit of time comparing the two sets of results. This could be used for time trends across years, or age patterns. The options for faceting or coloring allow the user to customize the plot to maximize utility. This function assumes that the two sets of estimates are in two different columns, e.g. old_mean and new_mean, and uses pivot_longer from tidyr to reshape the data long to make plotting more straightforward and flexible.

plot_trends(
  data,
  time_var,
  y1_var,
  y2_var,
  facet_x = NULL,
  facet_y = NULL,
  title = NULL,
  facet_type = "none",
  colors = c("dodgerblue", "deeppink4"),
  my_labs = c(y1_var, y2_var),
  line_size = 1
)

Arguments

data	data.table or data.frame. Data to plot comparing two sets of estimates, e.g. output of rank_change function. Required!
time_var	character() Variable name for time, e.g. year/year_id or ages. Required!
y1_var	character() Variable name for first set of results/estimates, e.g. the previous or old set for comparison. Required!
y2_var	character() Variable name for second set of results/estimates, e.g. the updated or new set for comparison. Required!
facet_x	character() Variable to facet by, along x-axis. This is required in order to use facet_wrap and facet_grid. Defaults to NULL.
facet_y	character() Variable to facet by, along y-axis, if desired. This is required in order to use facet_grid. Defaults to NULL.
title	character() Title of plot, if desired. Defaults to NULL.
facet_type	character() Choice of "none", "wrap", or "grid". Defaults to "none". The "none" option will create a single plot. The "wrap" option will produce a plot using facet_wrap. The "grid" option will produce a plot using facet_grid. Required!
colors	Vector. This vector would contain the name of two colors to be used in the plot to differentiate the two versions or sets of values. This defaults to dodgerblue and deeppink4, but can be otherwise specified by the user.
my_labs	Vector. This vector would contain the labels/names you want to use for the two different sets of estimates. If left blank this will default to the column names for y1_var and y2_var.
line_size	numeric() Can select line thickness. Default is 1.

Value

A ggplot object.

Examples

plot_trends(trend_changes_USA, 
            time_var = "year_id", 
            y1_var = "old_mean", 
            y2_var = "new_mean", 
            facet_x = "ihme_loc_id",
            my_labs = c("previous","new"),
            facet_type = "wrap",
            line_size = 2)