plots | R Documentation |
plot_forecasts_error_lead
: lots the raw error (estimate - observation) as a function of lead time across model runs from different forecast origins for multiple models and multiple species (or total) within a data set.
plot_covariates
: plots an observed timeseries and forecast timeseries of the covariates used.
plot_forecast_ts
: plots an observed timeseries and forecast timeseries with a prediction interval. Observations that occurred after the forecast are shown connected directly to the pre-cast observation data (as the black solid line with points).
plot_forecast_point
: plots the point value with confidence interval for a time point across multiple species. Casts can be selected either by supplying a forecast_id
number or any combination of dataset
, model
, and historic_end_newmoonnumber
, which filter the available forecasts in unison. This plot type can only handle output from a single forecast, so if multiple forecasts still remain, the one with the highest number is selected. To be more certain about forecast selection, use the forecast_id
input.
plot_forecasts_cov_RMSE
: plots the coverage (fraction of predictions within the CI) and RMSE (root mean squared error) of each model among multiple species.
plot_forecasts_error_lead(
main = ".",
forecasts_ids = NULL,
forecasts_evaluations = NULL,
historic_end_newmoonnumbers = NULL,
models = NULL,
datasets = NULL,
species = NULL
)
plot_forecasts_cov_RMSE(
main = ".",
forecasts_metadata = NULL,
forecasts_ids = NULL,
forecasts_evaluations = NULL,
historic_end_newmoonnumbers = NULL,
models = NULL,
datasets = NULL,
species = NULL
)
plot_forecast_point(
main = ".",
forecasts_metadata = NULL,
forecast_id = NULL,
dataset = NULL,
model = NULL,
historic_end_newmoonnumber = NULL,
species = NULL,
highlight_sp = NULL,
newmoonnumber = NULL,
with_census = FALSE
)
plot_forecast_ts(
main = ".",
forecasts_metadata = NULL,
forecast_id = NULL,
dataset = NULL,
model = NULL,
historic_start_newmoonnumber = NULL,
historic_end_newmoonnumber = NULL,
species = NULL
)
plot_covariates(main = ".", to_plot = "ndvi")
main |
|
forecasts_evaluations |
|
species |
|
forecasts_metadata |
|
forecast_id, forecasts_ids |
|
dataset, datasets |
|
model, models |
|
historic_end_newmoonnumber, historic_end_newmoonnumbers |
|
highlight_sp |
|
newmoonnumber |
|
with_census |
|
historic_start_newmoonnumber |
|
to_plot |
|
Casts can be selected either by supplying a forecast_id
number or any combination of dataset
, model
, and historic_end_newmoonnumber
, which filter the available forecasts in unison. This plot type can only handle output from a single forecast, so if multiple forecasts still remain, the one with the highest number is selected. To be more certain about forecast selection, use the forecast_id
input.
As of portalcasting v0.9.0
, the line and bands in plot_forecast_ts
and point and bars in plot_forecast_point
represent the mean and the 95 percent prediction interval.
NULL
. Plot is generated.
## Not run:
main1 <- file.path(tempdir(), "figures")
setup_production(main = main1)
plot_covariates(main = main1)
portalcast(main = main1, models = "AutoArima")
ids <- select_forecasts(main = main3,
species = c("DM", "PP", "total"),
models = c("AutoArima", "ESSS", "pevGARCH", "nbGARCH", "jags_RW"),
datasets = c("all", "controls"))$forecast_id
nids <- length(ids)
nsample_ids <- 1000
forecasts_ids <- ids[round(seq(1, nids, length.out = nsample_ids))]
evaluate_forecasts(main = main3,
forecasts_ids = forecasts_ids)
plot_forecast_ts(main = main1)
plot_forecast_point(main = main1)
plot_forecasts_error_lead(main = main1)
plot_forecasts_cov_RMSE(main = main1,
models = "AutoArima",
species = "DM")
unlink(main1, recursive = TRUE)
## End(Not run)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.