gocerry project.R

#Final update in Project
library(shiny)
library(ggplot2)
library(dplyr)# duplicated&& group by
library(arules)
library(readr)

# Data cleaning function
clean_data <- function(df) {
  # Remove duplicates
  df <- distinct(df)
  
  # Remove rows with any NA values
  df <- na.omit(df)
  
  # Check and convert data types
  for (col in names(df)) {
    if (col == "count" && !is.integer(df[[col]])) {
      df[[col]] <- as.integer(df[[col]])
    }
    if (col == "total" && !is.numeric(df[[col]])) {
      df[[col]] <- as.numeric(df[[col]])
    }
    if (col == "rnd" && !is.integer(df[[col]])) {
      df[[col]] <- as.integer(df[[col]])
    }
    if (col == "age" && !is.integer(df[[col]])) {
      df[[col]] <- as.integer(df[[col]])
    }
    if (col == "items" && !is.character(df[[col]])) {
      df[[col]] <- as.character(df[[col]])
    }
    if (col == "customer" && !is.character(df[[col]])) {
      df[[col]] <- as.character(df[[col]])
    }
    if (col == "city" && !is.character(df[[col]])) {
      df[[col]] <- as.character(df[[col]])
    }
  }
  return(df)  # This is cleaned Data
}

# Define UI-->  defines the appearance of the application 
ui <- fluidPage(#وهذا يعني ان العناصر ستتمد لملا الشاشه تلقائيا
  titlePanel("Dataset Analysis App"),#title of the program
  sidebarLayout(#بنظم الواجهه الي هتظهر لما ارن الكود
    sidebarPanel(
      # here we take the input
      fileInput("file_path", "Upload CSV File:", accept = c(".csv")),
      numericInput("numofcls", "Number of Groups (2-4):",value=2, min = 2, max = 4, step = 1),
      numericInput("minsupp", "Minimum Support:", value = 0.01, min = 0, max = 1, step = 0.01),
      numericInput("minconf", "Minimum Confidence:", value = 0.5, min = 0, max = 1, step = 0.01),
      actionButton("process", "Process Data"),
      
      
      actionButton("run_button", "Run Apriori"),
      width = 3
    ),
    mainPanel(#output appears here  
      tabsetPanel(
        tabPanel("Cleaned Data",  tableOutput("cleaned_data")),#panel:appears top of the program,when click here appears the cleaned data in table(table output)
        tabPanel("Clustered Data", plotOutput("plot")),#output as plot
        tabPanel("Clustered Data Summary", tableOutput("cluster_summary")),#output as table
        tabPanel("Visualization", 
                 selectInput(inputId= "tools", label = "Choose a graph to be displayed",
                             choices = c("Compare cash and credit totals." = "piechart",
                                         "Display each city total spending" = "barplot",
                                         "Comparing each age and sum of total spending"="plot",
                                         "Distribution of total spending"="boxplot",
                                         "Dashboard"="dashboard")),
                 plotOutput("visualization")),#output of this is->visualization graphs
        tabPanel("Apriori Association Rule Mining",
                 dataTableOutput("rules_table"),
                 plotOutput("relative_plot"),
                 plotOutput("absolute_plot")
        ))
    )#C:/Users/HP/OneDrive/Documents/Data_set_Project.csv
  )
)


# Define server logic
server <- function(input, output) {
  
  cleaned_data <- reactiveVal(NULL)
  df <- reactiveVal(NULL)
  
  observeEvent(input$process, {#processاضغط علي الزر الي عملته وهو 
    req(input$file_path)
    df_data <- read.csv(input$file_path$datapath)
    
    # Clean data
    df_cleaned <- clean_data(df_data)
    output$cleaned_data <- renderTable({#call front from ui 
      df_cleaned 
    })
    
    # Perform KMeans clustering
    kmeanfn <- function(df, nclusters) {
      processed_data <- df %>% #المتغير الي قبل العلامه بيحصل عليه التعديلات الي بعدالعلامه 
        group_by(customer,age) %>% #group by-> makes new data frame
        summarise(avg_total = sum(total))
      #Group by w summarize eletnen by3mlo  el dataframe ...w group by sh8ala 3la awl 2 columns w summarize sh8ala 3la total avg Group by btgm3 el duplicates w summerize bt sum el total bta3 ay 7aga etkrret w gm3naha b group by
      
      clusterdata <- data.frame(age = processed_data$age, total = processed_data$avg_total)
      cluster <- kmeans(clusterdata, centers = nclusters)
      newD <- cbind.data.frame(processed_data,  (cluster$cluster))#second cluster is default        lara&&mariam
      names(newD) <- c("name", "age", "total","Cluster_Group")
      return(newD)
    }
    
    clustered_data <- kmeanfn(df_cleaned, input$numofcls)# return our table with groups
    
    output$plot <- renderPlot({
      ggplot(clustered_data, aes(x = age, y = total, color = Cluster_Group)) + #بيقسم الالوان علي حسب عدد الجروبس
        geom_point(size=5) +#("+" is syntax)             
        labs(title = "KMeans Clustering Results")
    })
    
    output$cluster_summary <- renderTable({#lara????
      clustered_data
    })
    
    output$visualization <- renderPlot({
      if(input$tools=='barplot'){
        df_grp_city = df_cleaned %>% group_by(city)  %>%
          summarise(total_spending = sum(total), 
                    .groups = 'drop')
        df_grp_city = df_grp_city[order(df_grp_city$total_spending,decreasing=TRUE),]#mariam????decreasing
        barplot(height=df_grp_city$total_spending,
                name=df_grp_city$city,
                col="violetred3",
                main="display each city and its total spending",
                xlab="cities",
                ylab="total spending",
                las=2, #بتخليها بالعرض     cities
                cex.names = 0.7)# resize x axis names
      }
      else if(input$tools=="piechart"){
        x = table(df_cleaned$paymentType)#counting number of credit and cash
        percentage = paste0(round(100*x/sum(x)),"%")
        pie(x,percentage,main=" payment type percentage",col=c("purple","pink"))
        legend("bottomright",legend=c("cash","credit"),fill=c("purple","pink"))
      } else if(input$tools=="plot"){
        df_grp_age = df_cleaned %>% group_by(age)  %>%
          summarise(total_spending = sum(total), 
                    .groups = 'drop')
        ggplot(data=df_grp_age,mapping=aes(x=age,y=total_spending))+
          geom_point(size=5) + 
          geom_line(colour="hotpink")
      } else if(input$tools=="boxplot"){
        boxplot(x= df_cleaned$total , main="Distribution of total spending" , xlab="total spending",col="mediumaquamarine")
      }
      else if(input$tools=="dashboard"){
        par(mfrow=c(2,2))
        x = table(df_cleaned$paymentType)#counting number of credit and cash
        percentage = paste0(round(100*x/sum(x)),"%")
        pie(x,percentage,main=" payment type percentage",col=c("paleturquoise2","violetred3"))
        legend("bottomright",legend=c("cash","credit"),fill=c("paleturquoise2","violetred3"))
        #_____________________________________________________________________________________________________________________
        df_grp_age = df_cleaned %>% group_by(age)  %>%
          summarise(total_spending = sum(total), 
                    .groups = 'drop')
        plot(x=df_grp_age$age,y=df_grp_age$total_spending,
             main = "comparing each age with the sum of total spending",
             xlab="age",ylab = "total spending of a specific generation",col="maroon4")
        #_____________________________________________________________________________________________________________________
        boxplot(x= df_cleaned$total , main="Distribution of total spending" , xlab="total spending",col="paleturquoise2")
        #_____________________________________________________________________________________________________________________
        df_grp_city = df_cleaned %>% group_by(city)  %>%
          summarise(total_spending = sum(total), 
                    .groups = 'drop')
        df_grp_city = df_grp_city[order(df_grp_city$total_spending,decreasing = TRUE),]
        barplot(height=df_grp_city$total_spending,
                names=df_grp_city$city,
                col="violetred3",
                main="display each city and its total spending",
                xlab="cities",
                ylab="total spending",
                las=2, #بتخلي الارقام الموجوده علي ص بالعرض 
                cex.names = 0.4)# resize x axis names
        
      }}) 
    
    # Function to read and process the dataset for Apriori
    dataset <- eventReactive(input$run_button, {
      # df <- read.csv(input$file_path) #Read the file
      #df_cleaned <- clean_data(df()) 
      transaction <- strsplit(df_cleaned$items, ",")#الكلمه وكومه عشان تتحول لترانسيكشن
      trans <- as(transaction, "transactions")
      return(trans)
    })
    
    # Run Apriori and display rules
    apriori_rules <- eventReactive(input$run_button, {
      trans <- dataset()#اوصل للبيانات بسهوله  بيخليني اشوفها
      rules <- apriori(trans, parameter = list(supp = input$minsupp, conf = input$minconf, minlen = 2))
      return(rules)
    })
    
    
    # Output rules as a table
    output$rules_table <- renderDataTable({
      rules <- apriori_rules()
      inspect(rules)
    })
    
    # Plot relative item frequencies  (numerical)
    output$relative_plot <- renderPlot({
      trans <- dataset()
      itemFrequencyPlot(trans, topN = 5, type = "relative",col='lightblue1')
    })#اعلي 5 عليهم طلب اكتر
    
    
    # Plot absolute item frequencies
    output$absolute_plot <- renderPlot({
      trans <- dataset()
      itemFrequencyPlot(trans, topN = 5, type = "absolute",col="plum3")
    })
  })#
}

# Run the application
shinyApp(ui = ui, server=server)