In EBukin/devPTIpack: Development environment for Precise Index Targeting analysis

devPTIpack

The goal of the devPTIpack is to be a platform that contains key instruments for creating PTI dashboards as well as embedding PTI modules in to the existing shiny apps.

Installation:

# Install development version from GitHub
remotes::install_github("wbPTI/devPTIpack")

# Install specific version from GitHub
remotes::install_github("wbPTI/devPTIpack@v0.0.9")

We add @v0.0.9 to be able to refer to a specific tag. Check the tags' tree on github for existing tags.

Usage

There are several ways how this package can be used:

to deploy a stand-along PTI apps to the server;
to embed PTI modules into existing apps;

Run a stand-along PTI app

PTI apps are pre-packaged in a form of two functions.

devPTIpack::launch_pti()
devPTIpack::launch_pti_onepage()

Minimum data preparation requirements must be met in order to launch a PTI using functions from above. We need to have shp_dta, inp_dta and the app's name. In more details, data preparation is explained in vignette("dataprep"). Some exemplary data sets are included in the package: geometries devPTIpack::ukr_shp and metadata devPTIpack::ukr_mtdt_full.

# Running PTI app with sample data
library(devPTIpack)
launch_pti_onepage(
  shp_dta = ukr_shp,
  inp_dta = ukr_mtdt_full,
  app_name = "Sample PTI"
)

Many other parameters could be specified in devPTIpack::launch_pti(). The most notable are:

app_name - title of the app. Appears across multiple places in the app.
show_waiter - TRUE/FALSE indicates weather a "waiter" screen has to show, when the app launches.
show_adm_levels - vector c("admin1", "admin4") or NULL, indicates at what admin levels the map must be displayed If NULL all available administrative levels are used.
shapes_path and mtdtpdf_path are relative or absolute path to a zip archive with shape files used for mapping and metadata pdf respectively.
wt_dwnld_options and map_dwnld_options specify what data and metadata download options are available in different places of the app's interface.

Deploying a stand-along PTI app

When data is prepare properly and devPTIpack::launch_pti_onepage() works, we can create a dedicated folder, where app's data and code can be stored. This is needed if we want to deploy the PTI app to a Shiny server or an RStudio Connect server.

To create a new PTI project use devPTIpack::create_new_pti():

devPTIpack::create_new_pti("PATH-TO-NEW-PROJECT-LOCATION/PROJECT-NAME")

PATH-TO-NEW-PROJECT-LOCATION is a relative or an absolute path to the folder, where the app should be created.
a new sub-folder with the app itself will be created with the named PROJECT-NAME.

fs::dir_tree(here::here("inst/template_pti/"))

This is how a typical PTI folder looks like.

+-- app-data
|   \-- metadata.Rmd
+-- app.R
+-- app-name.Rproj
+-- landing-page.md
\-- R
|   \-- _disable_autoload.R

After the basic app folder was created, users need to populate it with data and make some minor edits.

Step 1. Prepare proper metadata.xlsx and shapes.rds files; Files names can be anything. One will have to specify correct paths to the files in app.R.
Step 2. Place shapes and metadata into the app-data folder. Use the app-data folder to store also other data such as archive with the geometrical boundaries or metadata pdf documentation.
Step 3. Knit metadata.Rmd in order to generate metadata.pdf documentation of the data used in the app. This documentation is based on the metadata.xlsx and shapes.rds. Remember to re-render metadata PDF whenever the metadata changes.
Step 4. Modify the app.R file specifying paths to metadata.xlsx, shapes.rds, pdf documentation and archive with the geometrical boundaries. Change the app's name and tune devPTIpack::launch_pti() function according to your needs.

Note, that inp_dta parameter in devPTIpack::launch_pti() shell be specified as a pre-loaded r object. As it is usually generated based on the metadata.xlsx, use devPTIpack::fct_template_reader() (see ?devPTIpack::fct_template_reader).

Step 5. Run the app and troubleshoot its functioning. Remember, if the app does not start, it is likely that there are some discrepancies in the data preparation. Always use the built-in data to check if the app launches from its own folder. You may try using devPTIpack::validate_geometries() or devPTIpack::validate_metadata() to check if geometries or metadata are prepared properly. These validation functions, however, are not exhaustive, thus, consult vignette("dataprep") or ask for help.
Step 6. Develop additional functionality in folder R.
Step 7. landing-page.md is an optional parameter and may be used for customizing the info modal that pops-up when the app launches.

Fully, completed app folder may look like this:

fs::dir_tree(here::here("../examplePTIapp/"))

+-- app-data
|   +-- admin_bounds.rds
|   +-- metadata.pdf
|   +-- metadata.Rmd
|   +-- mtdt.xlsx
|   \-- shapefiles.zip
+-- app.R
+-- examplePTIapp.Rproj
+-- landing-page.md
\-- R
    \-- _disable_autoload.R

The app.R script, which launches the app and could be used to deploy such app to an external server is:

# Launch the ShinyApp (Do not remove this comment)
# To deploy, run: rsconnect::deployApp()
# Or use the blue button on top of this file

library(shiny)
library(devPTIpack)

launch_pti(
  shp_dta = readRDS("app-data/admin_bounds.rds"),
  inp_dta = devPTIpack::fct_template_reader("app-data/mtdt.xlsx"),
  app_name = "Example country", 
  show_waiter = TRUE, 
  show_adm_levels = c("admin1", "admin2"),
  shapes_path = "app-data/shapefiles.zip", 
  mtdtpdf_path = "app-data/metadata.pdf",
  pti_landing_page = "./landing-page.md"
)

Embedding PTI modules into existing apps;

Essentially, PTI is a set of modules, which are used in different combinations. devPTIpack::launch_pti_onepage() offers simple glance at one key meta-module that could be used to embed PTI into any Shiny app.

# Server side
?devPTIpack::mod_ptipage_newsrv

# UI side
?devPTIpack::mod_ptipage_box_ui
?devPTIpack::mod_ptipage_twocol_ui

devPTIpack::mod_ptipage_newsrv() could be paid with any of two available UI and launched as an app in the command line. Note that in order for the module to function with its UI, they have to be identified in the same namespace provided with id parameter. For more information on modules, see 19 Shiny modules in Mastering Shiny or Modularizing Shiny app code or 10 Building the App with {golem}.

library(shiny)
library(devPTIpack)

# user interface
ui <- mod_ptipage_twocol_ui(id = "pti_mod")

# server
server <- function(input, output, session) {
  mod_ptipage_newsrv(id = "pti_mod",
                     inp_dta = reactive(ukr_mtdt_full),
                     shp_dta = reactive(ukr_shp), 
                     show_waiter = TRUE)
  }

# Shiny app.
shinyApp(ui, server)

Here is another example of embedding the PTI into an existing dashboard. It should be runnable in the console if all necessary packages are installed.

Note that most of the code below describe the dashboard layout and structure and only a few lines are dedicated to the PTI.
Remember to disable waiter when embedding module in a multitab dashboard.
More elaborate example of multitab dashboard is in the source code of the function devPTIpack::launch_pti.

## app.R ##
library(shiny)
library(shinydashboard)
library(devPTIpack)

ui <- dashboardPage(
  dashboardHeader(title = "Basic dashboard"),
  ## Sidebar content
  dashboardSidebar(
    sidebarMenu(
      menuItem("Dashboard", tabName = "dashboard", icon = icon("dashboard")),
      menuItem("PTI", tabName = "pti", icon = icon("th"))
    )
  ),
  ## Body content
  dashboardBody(
    tabItems(
      # First tab content
      tabItem(tabName = "dashboard",
        fluidRow(
          box(plotOutput("plot1", height = 250)),

          box(
            title = "Controls",
            sliderInput("slider", "Number of observations:", 1, 100, 50)
          )
        )
      ),

      # Second tab content
      tabItem(tabName = "pti",
        fluidRow(
          box(

            ### ### ### ### ### ### ## ### ### ### ### ###
            ### ### ### ### ### PTI UI ### ### ### ### ###
            ### ### ### ### ### ### ## ### ### ### ### ###
            devPTIpack::mod_ptipage_box_ui(
              id = "pti_mod", 
              map_height = "575px", 
              side_width = "325px",
              wt_style = "zoom:0.8;"
              ),
            width = 10,
            height = "600px"
            )
        )
      )
    )
  )
)

server <- function(input, output) {
  set.seed(122)
  histdata <- rnorm(500)

  output$plot1 <- renderPlot({
    data <- histdata[seq_len(input$slider)]
    hist(data)
  })


  ### ### ### ### ### ### ## ### ### ### ### ###
  ### ### ### ### ### PTI SERVER ### ### ### ###
  ### ### ### ### ### ### ## ### ### ### ### ###
  mod_ptipage_newsrv(
    id = "pti_mod",
    inp_dta = reactive(ukr_mtdt_full),
    shp_dta = reactive(ukr_shp), 
    show_waiter = TRUE
    )
  }

shinyApp(ui, server)