2. Adding model_viewer Shiny app

[gareth-j]

NOTE: #147 should go in before this PR.

• Summary of changes (Bug fix, feature, docs update, ...)

This adds a new Shiny app to view model output. It takes the output of a model run and lets the user plot a range of differnet plots and a prediction field over a leaflet map.

• Please check if the PR fulfills these requirements

• Closes Model output visualisation Shiny app #111
• Tests added and passed
• All code checks passing - styler run over code
• Vignettes updated/added
• Added an entry in the latest CHANGELOG.md file if fixing a bug or adding a new feature
• Any new libraries added to DESCRIPTION

[XueqingYin]

1. Thank you @gareth-j . I’ve test it on the covid data in Bristol, but the model_viewer app returns the following error:
   

This is because of the wrong argument name used when calling the create_prediction_field function within the functionmodel_viewer_shiny(). It should be
create_prediction_field (…, data_dist = data_type ,…), rather than create_prediction_field (…, data_type = data_type ,…) in shiny_modelviewer.R
Alternatively, this can be fixed by changing the argument name “data_dist” in create_prediction_field() in model.parse.R into “data_type”, i.e, create_prediction_field<- function (…, data_type = "Poisson"){ }

2. In the create_prediction_field() function in model.parse.R, it should be if (data_dist == "poisson") z <- base::exp(z)
   rather than if (data_type == "poisson") z <- base::exp(z)

3. I also checked the prior shiny app, but it gave an error:
   

Solution: I think we should use the following codes in shiny_priors.R to check and set the default CRS if it’s not given.

if ( (is.null(mesh_crs) || is.na(mesh_crs)) && (is.na(spatial_crs) || is.null(spatial_crs)) ) {
  warning("Cannot read CRS from mesh or spatial_data, using default CRS = +proj=longlat +datum=WGS84")
  crs <- "+proj=longlat +datum=WGS84"
} else if (is.na(mesh_crs) || is.null(mesh_crs) ) {
  crs <- spatial_crs
} else {
  crs <- mesh_crs
}

[XueqingYin]

Thanks @gareth-j, the errors are gone now.

Just a remainder that in the shiny_priors.R, the function launching the app is called interactive_priors() , but in the vignette priors_app.rmd, it's called fdmr::model_builder(...)

[gareth-j]

Thanks @gareth-j, the errors are gone now.

Just a remainder that in the shiny_priors.R, the function launching the app is called interactive_priors() , but in the vignette priors_app.rmd, it's called fdmr::model_builder(...)

I just noticed that and fixed it, thanks!