Shiny_1_Haiti_Event_Explorer_Map_Fatalities

#Shiny Project 1: Haiti 2018-2022: Conflict Event Explorer
#With pull down menu of event types, reactive leaflet map, and line graph of fatalities for corresponding sub_event_types


#TO DO: Fix scaling axes on fatality chart so that the years remain the same regardless of the event type choice.
#TO DO: Make better labels for each section. Add a text box to the sidebar with an explaination about what the app does.
#TO DO: Adjust theme to make it prettier.
#TO DO: Change markers on the map to points.
#TO DO: Fix event category Explosions/Remote Violence so that total fatalities displays correctly and the map stays on Haiti even on years with no incidents of that category.

#Required libraries: shiny, shinythemes, tidyverse, leaflet
library(shiny)
library(leaflet)
library(shinythemes)
library(tidyverse)


#Define the theme, physical layout of the page, and the user input
#Removed event category Explosions/Remote violence because n=2, and fatalities were not displaying correctly
ui <- fluidPage(
  titlePanel('Haiti 2018-2022: Conflict Event Explorer'),
  #shinythemes::shinytheme("superhero"), 
  sidebarLayout(
    sidebarPanel(
      selectInput(
        inputId = "event",
        label = "Select an event type",
        choices = c("Battles", "Protests", "Riots", "Violence against civilians"),
        selected = "Battles"
      ),
      selectInput(
        inputId = "year",
        label = "Select year",
        choices = c("2018", "2019", "2020", "2021", "2022"),
        selected = "2018"
      ),
      h4 ("Select from the pull down menu of event categories and years, to see where violent events have occurred, and how different sub-event-types compare in terms of fatalities."),
      div(
        h5 (("Data Source: "),tags$a(href="http://www.acleddata.com", "Armed Conflict Location & Event Data Project"), ("(accessed 2023-04-06)."))
      )
    ),
    mainPanel(
      leafletOutput(outputId = "map"),
      br(),
      br(),
      h4("Total Annual Fatalities by Sub-Event Type"),
      plotOutput(outputId = "lineplot")
    )
  )
)

#Define the data source, in this case ACLED.
server <- function(input, output) {
  data_url <- "https://www.dropbox.com/s/m62qxc07hae0r6m/ACLED_2018-01-01-2023-03-31-Caribbean-Haiti_CLEAN.csv?dl=1"
  data <- read.csv(data_url)
  
  #Filter the data so that only complete years are included in the analysis (2018-2022), and to make sure that latitude and longitude are not NULL and are numerical data.
  #Formatting the latitude and longitude here is just to make sure that read.csv interprets it the way I formatted it during the data preparation phase.   
  filtered_data <- reactive({
    data %>%
      filter(event_type == input$event) %>% 
      filter(event_date_year != 2023) %>%
      filter (event_date_year == input$year) %>%
      filter(latitude != "") %>%
      filter(longitude != "") %>%
      mutate(
        longitude = as.numeric(longitude),
        latitude = as.numeric(latitude)
      ) %>%
      group_by(event_date_year, sub_event_type) 
    
  })
  #Format the data for the line graph of fatalities, referring to the above filtered_data which includes a reactive expression for event type.   
  filtered_data_fatalities <- reactive({
    data %>%
      filter(event_type == input$event) %>% 
      filter(event_date_year != 2023) %>%
      filter(latitude != "") %>%
      filter(longitude != "") %>%
      mutate(
        longitude = as.numeric(longitude),
        latitude = as.numeric(latitude)
      ) %>%
      group_by(event_date_year, sub_event_type) 
  })
  
  #Plot the line graph of fatalities, referring to the above filtered_data which includes a reactive expression for event type.  
  output$lineplot <- renderPlot({
    filtered_data_fatalities <- filtered_data_fatalities()  #refers to the data which has responded to the user input
    fatalities_sum <- filtered_data_fatalities() %>%
      group_by(event_date_year, sub_event_type) %>%
      summarise(total_fatalities = sum(fatalities)) #calculates fatality sums and assigns them a name
    
    #Format the graph referring to the above data.
    ggplot(data = fatalities_sum) +
      geom_line(mapping = aes(x = event_date_year, y = total_fatalities , color = sub_event_type)) +
      labs(x = "Year", y = "Number of Fatalities") +
      xlim(2018, 2022)
  })
  
  #Format the map using the above data, directing it specifically to the data which has been filtered according to user input.  
  output$map <- renderLeaflet({
    leaflet() %>%
      addProviderTiles(providers$OpenStreetMap.HOT) %>% #choose tile source for basemap
      addMarkers(
        data = filtered_data(),
        lng = filtered_data()$longitude,
        lat = filtered_data()$latitude,
        popup = ~paste(strong("Location: "), filtered_data()$location, strong("  Sub-Event-Type: "), filtered_data()$sub_event_type, strong("  Date:"), filtered_data()$event_date)
      )
  })
}

#This activates the app
shinyApp(ui = ui, server = server)