In khufkens/cropmonitor: IFPRI cellphone based crop monitoring application

\begin{center} Report for farmer: # r params$n$uniqueuserid[1], field: # r params$n$uniquecropsiteid[1], max damage: r params$stats[2] \%, PBI payout: r params$stats[3], loss category: r params$stats[4], PBI reason: r params$stats[5] \end{center}

  plot_thumbs = function(df = params$n){
    label = c("a","b","c","d","e")
    loc = c(round(seq(1,nrow(df),nrow(df)/4)),nrow(df))
    par(mfrow=c(1,5),
        oma = c(5,0,0,0))
    for (i in 1:5){
      if (file.exists(df$thumbs[loc[i]])){
        r = raster::brick(df$thumbs[loc[i]])
        raster::plotRGB(r)
        text(350,250,label[i], cex = 1.5)
      }
    }
  }
  if (nrow(params$n) > 10){
    plot_thumbs()
  }

    library(ggplot2)
    library(ggthemes)

  if (nrow(params$n) > 10){
    # set values
    df = params$n
    span = 0.8

    # add labels
    label = c("a","b","c","d","e")
    loc = c(round(seq(1,nrow(df),nrow(df)/4)),nrow(df))
    df$label[loc] = label
    labels = df$label

    # sort things, database isn't ordered  
    gcc = df$gcc_90
    date = df$datetime
    thumbs = df$thumbs
    full_date = seq(min(date,na.rm = TRUE),
                    max(date,na.rm = TRUE),
                    'days')

    # smooth the data using a loess fit
    fit = loess(gcc ~ as.numeric(date), span = 0.3)
    fit_gcc = predict(fit, as.numeric(date), se = TRUE)
    gcc_smooth = fit_gcc$fit
    ci_up = gcc_smooth + fit_gcc$se * 1.96
    ci_down = gcc_smooth - fit_gcc$se * 1.96
    df = data.frame(date,gcc_smooth,ci_up,ci_down, gcc, labels)

    p = ggplot(df, aes(date,gcc_smooth)) + 
    ylim(low = min(ci_down, na.rm = TRUE),
         high = max(gcc, na.rm = TRUE) + 0.03) +
      xlab("") +
      ylab("Greenness") +
    geom_line() +
    geom_ribbon(aes(ymin=ci_down, ymax=ci_up), alpha=0.2) +
    geom_point(aes(date,gcc), na.rm=TRUE) +
    geom_text(aes(date,max(gcc, na.rm = TRUE), label = labels, vjust = -0.5),
            color = "black",
            size = 5,
            na.rm=TRUE)  
    p = p + theme_economist() +
    scale_color_economist() +
    ggtitle("Vegetation Greenness")
    plot(p)
  }

    library(ggplot2)
    library(ggthemes)

  if (nrow(params$n) > 10){

    # set values
    df = params$n
    span = 0.8

    # sort things, database isn't ordered  
    gcc = df$sobel
    date = df$datetime
    thumbs = df$thumbs
    full_date = seq(min(date,na.rm = TRUE),
                    max(date,na.rm = TRUE),
                    'days')

    # smooth the data using a loess fit
    fit = loess(gcc ~ as.numeric(date), span = 0.3)
    fit_gcc = predict(fit, as.numeric(date), se = TRUE)
    gcc_smooth = fit_gcc$fit
    ci_up = gcc_smooth + fit_gcc$se * 1.96
    ci_down = gcc_smooth - fit_gcc$se * 1.96
    df = data.frame(date,gcc_smooth,ci_up,ci_down, gcc, labels)

    p = ggplot(df, aes(date,gcc_smooth)) + 
    ylim(low = min(ci_down, na.rm = TRUE),
         high = max(gcc, na.rm = TRUE) + 0.03) +
      xlab("") +
      ylab("Texture metric") +
    geom_line() +
    geom_ribbon(aes(ymin=ci_down, ymax=ci_up), alpha=0.2) +
    geom_point(aes(date,gcc), na.rm=TRUE)

    p = p + theme_economist() +
    scale_color_economist() +
    ggtitle("Vegetation Texture")
    plot(p)
  }

This is a summary overview of the pictures taken during the growing season of r sprintf("%s - %s",min(format(df$date, "%Y"),na.rm = TRUE), max(format(df$date, "%Y"), na.rm = TRUE)).

khufkens/cropmonitor documentation built on May 31, 2019, 8:29 a.m.