I think it would be nice to have a possibility to plot input data to see how they look before passing them into a gadget model. I have some functions in my ghl model project which I could add. Now the question is how to name them to separate them from model output figures?

Output figures are currently named as plot_*. Would dplot_ (data plot) work? Should output figures then be renamed mplot_* or should we keep them as they are for consistency with the rgadget package?

Hi,

I thought it would be nice to be able to automatically plot a gadget model. Thought of basing this on g3_fit object. This is as far as I got for now. Posting it here in case someone gets inspired to continue. If not, I'll continue next year.

library(dplyr)
library(DiagrammeR)

stocks <- fit$res.by.year %>%
  filter(year == max(year)) %>%
  mutate(size = total.biomass/sum(total.biomass),
         type = "stock") %>%
  rename("name" = "stock") %>%
  dplyr::select(name, type, size)

fleets <- fit$fleet.info %>%
  filter(year == max(year)) %>%
  mutate(size = amount/sum(amount),
         type = "fleet") %>%
  rename("name" = "fleet") %>%
  dplyr::select(name, type, size)

predation <- optim_fit$suitability %>%
  dplyr::ungroup() %>%
  dplyr::filter(suit > 0) %>%
  dplyr::select(stock, fleet) %>%
  dplyr::distinct()

graph <- create_graph()

## ####

graph %>%
  add_nodes_from_table(
    table = bind_rows(stocks, fleets) %>% rename("value" = "size"),
    label_col = name,
    type_col = type
  ) %>%
  add_edges_from_table(
    table = predation,
    from_col = stock,
    to_col = fleet,
    from_to_map = label
  ) %>%
  set_node_attrs(
    node_attr = 'fontcolor',
    values = 'black'
  ) %>%
  select_nodes(type == "stock") %>%

  set_node_attrs_ws(
    node_attr = fillcolor,
    value = "orange") %>%
  set_node_attrs_ws(
    node_attr = shape,
    value = "rectangle") %>%
  set_node_attrs_ws(
    node_attr = y,
    value = 4) %>%
  clear_selection() %>%
  render_graph("tree")

A list of suggestions for an 'Iceland' template (f98aa81):

1. Annual plots - F instead of HR
2. Maturity observations as points

It would be useful to feed additional plots, for instance, retro, bootstrap and jitter plots, into gadget_plots (file_type = 'html'). A quick fix for retros was done here https://github.com/willbutler42/gadgetplots2/blob/master/R/gadget_plots_GR.R

Possibly something like gadget_plots(fit, file_type = 'html', additional_plots = list(Retro = retroplot, Bootstrap = bootplot)) where retroplot and bootplot are plot objects

If there is lots of data, it can be hard to discern how well the model fits ldists and age-length data. Need to have a think how visualising this can be improved, maybe having a row limit per page (creating multiple pages if the limit is exceeded) or something along these lines would work

We should develop some plotting tools to facilitate model comparisons. A good starting point would be allowing each panel of the annual plots to include multiple fits. plot_biomass and plot_rec can do this using facet_wrap so just need to extend this capability to single panels. Would be handy for SIs too