Our Storyboard

1. Storyboard

Storyboard aims to visually maps out user’s experience. It is a tool for making strong visual connection between the insights uncovered based on research and user’s interaction with the R Shiny dashboard application. The interactive components and UI design aims to facilitate data (and geospatial) exploration and analysis for users to develop effective counter measures and strategies.

The prototype can be broadly classify into three key areas:

Aspatial Analysis
Geospatial Analysis
Confirmatory Analysis

And the team has conceptualised the proposed layouts and UI features as follows:

For enhanced user experience, the prototype included ‘filter’ components (i.e. parameter selections) and ‘chart interpretation’ boxes, and have aligned them mainly to the left side of the web pages. The ‘chart interpretation’ box provides brief explanation of how each chart can be interpreted.

Section One - Aspatial Analysis

Aspatial Analysis - Overview

This tab serves as the “landing page” that displays the map of Myanmar and its spatial points of armed conflicts over the years (i.e. 2010 to 2023). Figure below shows the UI interactive features in the Overview sub-tab.

Code chunk below shows the simplified version of UI and Server components in R Shiny application for Overview sub-tab.

Show code

# UI Components
AspatialOverviewrow1 <-  fluidRow(
  highchartOutput(), # display line chart
  sliderInput(), # select year range
  selectizeInput(), # select event (allows multiple selection)
  radioButtons(), # select administrative region level
  selectizeInput(), # select administrative region (allows multiple selection)
  actionButton(), # action button
  checkboxInput(), # conditional panel (option to display line chart)
  highchartOutput(), # display line chart
  checkboxInput(), # conditional panel (option to display map information)
  leafletOutput(), # display point spatial map
)

AspatialOverviewrow2 <-  fluidRow(
  DT::dataTableOutput() # display datatable
)

# Server Components
output1-1-1 <- renderLeaflet({}) # point spatial map
output1-1-2 <- renderLeHighchart({}) # line chart
output1-1-3 <- DT::renderDataTable({}) # datatable

Aspatial Analysis - Distribution Analysis

This analysis page allows users to perform data visualisation to show the armed conflict incidents’ (and fatalities) spatial and density distribution represented in choropleth map and density ridge plot respectively. Figure below shows the UI interactive features in the Distribution Analysis sub-tab.

Code chunk below shows the simplified version of UI and Server components in R Shiny application for Distribution Analysis sub-tab.

Show code

# UI Components
AspatialDistributionrow1 <-  fluidRow(
  sliderInput(), # select year range
  radioButtons(), # select administrative region level
  radioButtons(), # select display rates
  selectInput(), # select colour palette
  checkboxInput(), # Customise Spatial Map
  selectInput(), # Spatial Map Classification Type
  sliderInput(), # Number of Classes
  checkboxInput(), # Customise Density Ridge Plot
  selectInput(), # Density Ridge Style
  tmapOutput(), # choropleth map
  plotOutput() # density ridge plot
)

AspatialDistributionrow2 <-  fluidRow(
  textOutput()
)


# Server Components
output1-2-1 <- renderTmap() # display choropleth map
output1-2-2 <- renderPlot() # display density ridge plot

Section Two - Geospatial Analysis

Geospatial Analysis - Local Measures of Spatial Autocorrelation

This analysis page allows users to perform a cluster and outlier analysis, to identify significant clusters of high and low values and outliers. Using the Local Moran’s I statistic, features are categorised into 2 clusters (High-High, Low-Low), 2 outliers (High-Low, Low-High) and 1 insignificant classes.

Code chunk below shows the simplified version of UI and Server components in R Shiny application for Local Measures of Spatial Autocorrelation sub-tab.

Show code

# UI Components
Cluster2 <- fluidRow(
  selectInput(), # select period
  selectInput(), # select Moran Event Type
  radioButtons(), # select Contiguity method 
  selectInput(), # select Spatial Weights Style
  sliderInput(), # Number of Simulations
  actionButton(), # action button 
  radioButtons(), # select confidence interval 
  selectInput(), # select Lisa Classification
  selectInput(), # select Local Moran's Stat
  plotOutput(), # LocalMoranMap
  plotOutput(), # LisaMap 
  textOutput(),
  DT::dataTableOutput()
)

# Server Components
output2-1-1 <- renderPlot()
output2-1-2 <- renderPlot()
output2-1-3 <- renderDataTable()

Geospatial Analysis - Hot & Cold Analysis

This analysis page allows users to perform a Hot & Cold Spot analysis, to identify significant areas of high and low values based on a calculated distance. Using the Getis-Ord Gi* statistic, features are grouped together when simila High (Hot) or Low (Cold) values are found in a cluster.

Code chunk below shows the simplified version of UI and Server components in R Shiny application for Hot & Cold Analysis sub-tab.

Show code

# UI Components
HotCold1 <- fluidRow(
  selectInput(), # select period
  selectInput(), # select Moran Event Type
  radioButtons(), # select Contiguity method 
  sliderInput(), # Number of Simulations
  actionButton(), # action button 
  radioButtons(), # select confidence interval 
  selectInput(), # select local Gi stats
  plotOutput(), # GI Stats
  plotOutput(), # Sig hot/cold
  textOutput(),
  DT::dataTableOutput()
)

# Server Components
output2-2-1 <- renderPlot()
output2-2-2 <- renderPlot()
output2-2-3 <- renderDataTable()

Geospatial Analysis - Emerging Hot Spot Analysis

This analysis page allows users to perform an Emerging Hot Spot analysis, to reveal and describe how hot spots and cold spots have changed over time. After identifying temporal trends, features are classified into one of 17 ESRI hot spot classifications.

Code chunk below shows the simplified version of UI and Server components in R Shiny application for Emerging Hot Spot Analysis sub-tab.

Show code

# UI Components
EHSA2 <- fluidRow(
  selectInput(), # select event type
  radioButtons(), # select Contiguity Method"
  sliderInput(), # select Time Lag of spatial neighbours
  sliderInput(), # Number of Simulations
  actionButton(),
  checkboxInput(), # checkbox to show EHSA classes
  plotlyOutput(), # display EHSA ba chart 
  checkboxInput(), # checkbox to show GI* trend plot"
  selectizeInput(), # select district
  plotlyOutput(), # display Giplot
  radioButtons(), # select Confidence level
  plotOutput(), # EHSAmap
  textOutput(), # EHSAText
  DT::dataTableOutput(), # data table
  textOutput()
) 
# Server Components
output2-3-1 <- renderPlot()
output2-3-2 <- renderPlot()
output2-3-3 <- renderDataTable()

Section Three - Confirmatory Analysis

Confirmatory Analysis - One-Way ANOVA Test

This analysis page allows users to perform a One-Way ANOVA test to identify if there is any significant difference between the mean or median value for event types and the number of fatalities. If the p-value is below the critical value, it means that the null hypothesis has sufficient statistical evidence to support. Whereas if the p-value is above the critical value, the null hypothesis will be rejected due to insufficient statistical evidence.

Code chunk below shows the simplified version of UI and Server components in R Shiny application for One-way ANOVA Test sub-tab.

Show code

# UI Components
Confirm1 <- fluidRow(
  selectInput(), # select time
  selectizeInput(), # select event type
  actionButton(), # reset button
  selectInput(), # select test type
  selectInput(), # Pairwise Display
  selectInput(), # P-value adjustment method 
  radioButtons(), # select Confidence level
  plotOutput(), # display Anovaplot
)

Confirm2 <- fluidRow(
  textOutput()
)

# Server Components
output3-1-1 <- renderPlot()

Confirmatory Analysis - Visualising Categorical Data

This analysis page allows users to visualise categorical data using mosaic plot to find association between the variables. The size of each tile would represent the proportion of observations for the variable. The colour of each tile would represent the residual where red tiles indicate significant negative residual where frequency is less than expected and blue tiles indicate significant positive residual where frequency is more than expected. The intensity of the colour represents the magnitude of the residuals which is shown on the legend on the right.

Code chunk below shows the simplified version of UI and Server components in R Shiny application for Visualising Categorical Data sub-tab.

Show code

# UI Components
Confirm3 <- fluidRow(
  selectInput(), # select Year
  selectizeInput(), # Region
  selectizeInput(), # Event Type
  selectInput(), # Include Data with or without Fatalities
  actionButton(), 
  plotOutput() # plot Mosaic plot
)

Confirm4 <- fluidRow(
  textOutput()
)

# Server Components
output3-2-1 <- renderPlot()

2. R Shiny Application (simplified code)

The storyboard (in Section 1) facilitates the development of a prototype in R Shiny Application. Iterative prototyping will allow continuous improvement of the final project.

The proposed layouts and UI features

Code chunk below shows the simplified version of R Shiny Application for Decoding Chaos prototype.

#========================================================== 
## load R packages
#========================================================== 
pacman::p_load(shiny, shinydashboard, shinycssloaders, 
               tidyverse, dplyr, leaflet, leaflet.extras, plotly, 
               ggthemes, fresh, sf, sfdep, tmap, tm, 
               ggraph, DT, spatstat,
               lubridate,viridis, ggplot2, readr, purrr, ggstatsplot, 
               vcd, ggmosaic, forcats,
               ggridges, ggdist, highcharter)


#========================================================== 
## UI Components
#========================================================== 
# main header ---
header <- dashboardHeader(title = "Decoding Chaos")


# main sidebar ---
sidebar <- dashboardSidebar()

sidebarMenu(
    menuItem("Aspatial", tabName = "Aspatial"),
    menuItem("Geospatial", tabName = "Geospatial"),
    menuItem("Confirmatory", tabName = "Confirmatory"),
    menuItem("Visit ACLED data"))

# main body ---
body <- dashboardBody(
  tabItems(
    tabItem(tabName = "Aspatial",
            AspatialSubTabs
    ),
    tabItem(tabName = "Cluster",
            ClusterSubTabs 
    ),
    tabItem(tabName = "ConfirmatoryAnalysis",
            ConfirmSubTabs
)))

# fluidRows ---

# aspatial analysis tab
AspatialOverviewrow1 <-  fluidRow(
  highchartOutput(), # display line chart
  sliderInput(), # select year range
  selectizeInput(), # select event (allows multiple selection)
  radioButtons(), # select administrative region level
  selectizeInput(), # select administrative region (allows multiple selection)
  actionButton(), # action button
  checkboxInput(), # conditional panel (option to display line chart)
  highchartOutput(), # display line chart
  checkboxInput(), # conditional panel (option to display map information)
  leafletOutput(), # display point spatial map
)

AspatialOverviewrow2 <-  fluidRow(
  DT::dataTableOutput() # display datatable
)

AspatialDistributionrow1 <-  fluidRow(
  sliderInput(), # select year range
  radioButtons(), # select administrative region level
  radioButtons(), # select display rates
  selectInput(), # select colour palette
  checkboxInput(), # Customise Spatial Map
  selectInput(), # Spatial Map Classification Type
  sliderInput(), # Number of Classes
  checkboxInput(), # Customise Density Ridge Plot
  selectInput(), # Density Ridge Style
  tmapOutput(), # choropleth map
  plotOutput() # density ridge plot
)

AspatialDistributionrow2 <-  fluidRow(
  textOutput()
)

# geospatial analysis tab
Cluster2 <- fluidRow(
  selectInput(), # select period
  selectInput(), # select Moran Event Type
  radioButtons(), # select Contiguity method 
  selectInput(), # select Spatial Weights Style
  sliderInput(), # Number of Simulations
  actionButton(), # action button 
  radioButtons(), # select confidence interval 
  selectInput(), # select Lisa Classification
  selectInput(), # select Local Moran's Stat
  plotOutput(), # LocalMoranMap
  plotOutput(), # LisaMap 
  textOutput(),
  DT::dataTableOutput()
)


HotCold1 <- fluidRow(
  selectInput(), # select period
  selectInput(), # select Moran Event Type
  radioButtons(), # select Contiguity method 
  sliderInput(), # Number of Simulations
  actionButton(), # action button 
  radioButtons(), # select confidence interval 
  selectInput(), # select local Gi stats
  plotOutput(), # GI Stats
  plotOutput(), # Sig hot/cold
  textOutput(),
  DT::dataTableOutput()
)

EHSA2 <- fluidRow(
  selectInput(), # select event type
  radioButtons(), # select Contiguity Method"
  sliderInput(), # select Time Lag of spatial neighbours
  sliderInput(), # Number of Simulations
  actionButton(),
  checkboxInput(), # checkbox to show EHSA classes
  plotlyOutput(), # display EHSA ba chart 
  checkboxInput(), # checkbox to show GI* trend plot"
  selectizeInput(), # select district
  plotlyOutput(), # display Giplot
  radioButtons(), # select Confidence level
  plotOutput(), # EHSAmap
  textOutput(), # EHSAText
  DT::dataTableOutput(), # data table
  textOutput()
) 

# confirmatory analysis tab
Confirm1 <- fluidRow(
  selectInput(), # select time
  selectizeInput(), # select event type
  actionButton(), # reset button
  selectInput(), # select test type
  selectInput(), # Pairwise Display
  selectInput(), # P-value adjustment method 
  radioButtons(), # select Confidence level
  plotOutput(), # display Anovaplot
)

Confirm2 <- fluidRow(
  textOutput()
)


Confirm3 <- fluidRow(
  selectInput(), # select Year
  selectizeInput(), # Region
  selectizeInput(), # Event Type
  selectInput(), # Include Data with or without Fatalities
  actionButton(), 
  plotOutput() # plot Mosaic plot
)

Confirm4 <- fluidRow(
  textOutput()
)


# subtabs
AspatialSubTabs <- tabsetPanel(
  tabPanel("Overview", 
           AspatialOverviewrow1,
           AspatialOverviewrow2
  ),
  tabPanel("Distribution Analysis", 
           AspatialDistributionrow1,
           AspatialDistributionrow2
  )
)

ClusterSubTabs <- tabsetPanel(
  
  tabPanel("Local Measures of Spatial Autocorrelation", 
           Cluster2),
  tabPanel("Hot & Cold Spot Analysis(HCSA)", 
           HotCold1),
  tabPanel("Emerging Hot Spot Analysis", 
           EHSA2)
)

ConfirmSubTabs <- tabsetPanel(
  tabPanel("One-Way Anova Test", 
           Confirm1,
           Confirm2),
  #tabPanel("Mosaic Plot",
  #         Confirm2),
  tabPanel("Mosaic Plot-VCD",
           Confirm3,
           Confirm4)
  #Confirm2)
  
)


#========================================================== 
## UI dashboard
#========================================================== 
ui <- dashboardPage(title = 'Armed Conflicts in Myanmar (2010 to 2023)', 
                    header, sidebar, body)  


#========================================================== 
## Server Components
#========================================================== 
server <- function(input, output) {

# aspatial analysis tab
output1-1-1 <- renderLeaflet({}) 
output1-1-2 <- renderLeHighchart({}) 
output1-1-3 <- DT::renderDataTable({}) 

output1-2-1 <- renderTmap()
output1-2-2 <- renderPlot()

# geospatial analysis tab
output2-1-1 <- renderPlot()
output2-1-2 <- renderPlot()
output2-1-3 <- renderDataTable()

output2-2-1 <- renderPlot()
output2-2-2 <- renderPlot()
output2-2-3 <- renderDataTable()

output2-3-1 <- renderPlot()
output2-3-2 <- renderPlot()
output2-3-3 <- renderDataTable()

# confirmatory analysis tab
output3-1-1 <- renderPlot()
output3-2-1 <- renderPlot()
}


#========================================================== 
## Run Shiny Application
#========================================================== 
shinyApp(ui = ui, server = server)

1. Storyboard

Section One - Aspatial Analysis

Aspatial Analysis - Overview

Aspatial Analysis - Distribution Analysis

Section Two - Geospatial Analysis

Geospatial Analysis - Local Measures of Spatial Autocorrelation

Geospatial Analysis - Hot & Cold Analysis

Geospatial Analysis - Emerging Hot Spot Analysis

Section Three - Confirmatory Analysis

Confirmatory Analysis - One-Way ANOVA Test

Confirmatory Analysis - Visualising Categorical Data

2. R Shiny Application (simplified code)

Reference