Corrected Ngram Viewer

global.R

list.of.packages <- c("ggplot2", "devtools", "lubridate", "nlme", 
  "plotly", "shinydashboard", "shiny", "git2r")
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[, 
  "Package"])]
if (length(new.packages)) install.packages(new.packages)

library(shiny)
library(shinydashboard)

library(devtools)
install_github("ngramr", "seancarmody")
require(ngramr)
library(ggplot2)
library(lubridate)
library(nlme)
library(plotly)
library(shinydashboard)
library(shiny)
library(git2r)

server.R

# Setup Shiny app back-end components
# -------------------------------------
server <- function(input, output) {
  
  x.ng <- reactive({
    ng <- ngram(input$searchButton2, year_start = input$daterange[1], 
      year_end = input$daterange[2], count = TRUE, corpus = input$checkbox, 
      smoothing = 0)
    
  })
  x.etall <- reactive({
    
    ngram("et al", year_start = input$daterange[1], year_end = input$daterange[2], 
      count = TRUE, corpus = input$checkbox, smoothing = 0)
  })
  x.ofthe <- reactive({
    ngram("of the", year_start = input$daterange[1], year_end = input$daterange[2], 
      count = TRUE, corpus = input$checkbox, smoothing = 0)
  })
  x.the <- reactive({
    ngram("the", year_start = input$daterange[1], year_end = input$daterange[2], 
      count = TRUE, corpus = input$checkbox, smoothing = 0)
  })
  x <- reactive({
    df = data.frame(p.t = x.ng()$Count/x.the()$Count, q.t = x.etall()$Count/x.ofthe()$Count, 
      time = x.ng()$Year)
  })
  
  x.final <- reactive({
    fit = gls(p.t ~ time * q.t, data = x(), corr = corAR1(form = ~1))
    
    
    
    # ## simple examples using gamm as alternative to gam
    beta <- coef(fit)
    beta.trans <- c(beta[3] + beta[1], beta[4] + beta[2], 
      beta[1], beta[2])
    p.t_q.t <- pmax((beta.trans[1] + beta.trans[2] * x()$time) * 
      x()$q.t, 0)
    p.t_Nq.t <- pmax((beta.trans[3] + beta.trans[4] * x()$time) * 
      (1 - x()$q.t), 0)
    prop = c(p.t_q.t, p.t_Nq.t, x()$p.t)
    corpus = c(rep("Scientific literature", length(p.t_q.t)), 
      rep("Non scientific literature", length(p.t_Nq.t)), 
      rep("Google Viewer", length(x()$p.t)))
    df2 <- data.frame(proportion = prop * 100, Trends = corpus, 
      Year = x()$time)
    df2
    
  })
  
  plot.choice = reactive(input$plotChoice)
  
  output$plot.corrected <- renderPlotly({
    if (plot.choice() == "all") {
      df = x.final()
      title.choix <- "Google Viewer and corrected trend"
    }
    if (plot.choice() == "google") {
      df = subset(x.final(), Trends == "Google Viewer")
      title.choix <- "Google Viewer trend"
    }
    if (plot.choice() == "corrected") {
      df = subset(x.final(), Trends != "Google Viewer")
      title.choix <- "Corrected trend"
    }
    
    p2 <- ggplot(df, aes(x = Year, y = proportion, colour = Trends)) + 
      geom_line() + ggtitle(paste0(title.choix, " of ", 
      input$searchButton2, sep = " ")) + ylab("Proportion (%)")
    ggplotly(p2)
    
  })
  
  
}

ui.R

# This is the user-interface definition of a Shiny web application.
# You can find out more about building applications with Shiny here:
#
# http://shiny.rstudio.com
#


sidebar <- dashboardSidebar(
  textInput("searchButton2", h3("Word"), value= "Figure"),
  sliderInput("daterange", h3("Year range"),
              min = 1800, max = 2008, value = c(1900,2000)),
 # dateRangeInput("daterange", h3("Year range:"),format = "yyyy", start="1900-01-01", end="2000-01-01", startview = "decade"),
  selectInput("checkbox", 
                     label = h3("Choose the corpus"), 
                     choices = list("American English 2012" = "eng_us_2012", 
                                    "American English 2009" = "eng_us_2009",
                                   	"British English 2012" = "eng_gb_2012",
                                    	"British English 2009" = "eng_gb_2009",
                                    	"English 2012" = "eng_2012",
                                    	"English 2009" = "eng_2009",
                                    	"English Fiction 2012" = "eng_fiction_2012",
                                    	"English Fiction 2009" = "eng_fiction_2009",
                                    	"Google One Million" = "eng_1m_2009",
                                    	"French 2012" = "fre_2012",
                                    	"French 2009" = "fre_2009"
                                    ),
                     selected = "eng_us_2009")
  )



# Simple header -----------------------------------------------------------
dashboardPage(
  dashboardHeader(title = "Corrected Ngrams Viewer by Aurélien Nicosia", titleWidth = 450),
  dashboardSidebar(sidebar ,titleWidth = 350),
  dashboardBody(
    # Boxes need to be put in a row (or column)
    fluidRow(
           box(
             title = "What do you want to plot?",width= 8,collapsible= TRUE,status = "warning", solidHeader = TRUE, br(), 
             selectInput("plotChoice", 
                         label = NULL,
                         choices = list("Google Viewer and corrected trend" = "all", 
                                        "Google Viewer" = "google", "Corrected trend" = "corrected"),
                         selected = "all"
                         )),
             
             box(title= "Extracting useful information from the Google Ngram
dataset: A general method to take the growth of the scientific literature into account" ,statut = "primary",collapsible= TRUE,solidHeader = TRUE, br(),
                 plotlyOutput("plot.corrected"), width =8)
      
           
           
     )
    )
  )