Skip to contents

Create heatmap of timestep grid

grid_heatmap.Rd

Creates heatmap of input data frame values where rows are timesteps and columns are sites to visualize dataset scale correlation structure and missingness patterns. Can z-score standardize values by site to visualize relative temporal patterns in dataset.

Usage

grid_heatmap(
  grid,
  data_thresh = NA,
  order = "completeness",
  site_order = NA,
  zscore = T,
  rv_cb = T,
  cb_rng = NA,
  x_axis_breaks = 5,
  rnd = 3
)

Arguments

grid

data frame where rows and columns represent timesteps and sites, respectively. Accepts output from timestep_grid and imputed_grid in output list of impute_grid.

data_thresh

site completeness threshold. Only visualize columns (sites) with less than this proportion of observations (non-NA - not assigned - values) in constituent rows (timesteps). Default is NA (show all sites).

order

character string specifying site plotting order. Can be set to "completeness" (order sites by proportion complete), "cluster" (order sites by hierarchical agglomerative clustering), or "custom" (supply custom site order as vector using site_order argument). Default is "completeness".

site_order

character vector describing custom site plotting order. Automatically removes leading "X." prefixes to site identifiers if present. Default is NA.

zscore

logical flag to z-score standardize observed values by site before plotting. Default is TRUE.

rv_cb

logical flag to reverse colorbar scale. Default is FALSE.

cb_rng

two element numeric vector setting colorbar range. If values fall outside the specified range, colors are saturated at extremes when plotted. Default is NA, which autoscales colorbar range to the data.

x_axis_breaks

number of tick marks and labels to include on x-axis. Default is 5.

rnd

number of digits to round values to for plotting on heatmap. Default is 3. Set to NA to suppress rounding.

Value

named list containing:

grid_plot

ggplot object created using geom_raster.

site_order

character vector of site plotting order.

See also

geom_raster, hclust, timestep_grid, impute_grid

Author

Maintainer: Zeno F. Levy zlevy@usgs.gov

Examples

# load example Long Island dataset
  data(LI_data)

# grid data at monthly timestep using median observed values
  grid <- timestep_grid(data = LI_data, 
                        timestep = "monthly", 
                        agg_method = "median")

# visualize heatmap of timestep grid 
# only show sites that are at least 35 percent complete
  hm <- grid_heatmap(grid, data_thresh = 0.35)
  hm$grid_plot