pkg/R/plot.R

   1 #' Plot curves
   2 #'
   3 #' Plot a range of curves in data.
   4 #'
   5 #' @inheritParams computeError
   6 #' @param indices Range of indices (integers or dates)
   7 #'
   8 #' @export
   9 plotCurves <- function(data, indices=seq_len(data$getSize()))
  10 {
  11     series = data$getSeries(indices)
  12     yrange = quantile(series, probs=c(0.025,0.975), na.rm=TRUE)
  13     par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
  14     matplot(series, type="l", ylim=yrange, xlab="Time (hours)", ylab="PM10")
  15 }
  16
  17 #' Plot error
  18 #'
  19 #' Draw error graphs, potentially from several runs of \code{computeForecast()}.
  20 #'
  21 #' @param err Error as returned by \code{computeError()}
  22 #' @param cols Colors for each error (default: 1,2,3,...)
  23 <<<<<<< HEAD
  24 =======
  25 #' @param agg Aggregation level ("day", "week" or "month")
  26 >>>>>>> 3a38473a435221cc15b6215b0d6677cd370dc2d6
  27 #'
  28 #' @seealso \code{\link{plotCurves}}, \code{\link{plotPredReal}},
  29 #'   \code{\link{plotSimils}}, \code{\link{plotFbox}}, \code{\link{computeFilaments}},
  30 #'   \code{\link{plotFilamentsBox}}, \code{\link{plotRelVar}}
  31 #'
  32 #' @export
  33 plotError <- function(err, cols=seq_along(err), agg="day")
  34 {
  35     if (!is.null(err$abs))
  36         err = list(err)
  37     par(mfrow=c(2,2), mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
  38     L = length(err)
  39
  40     yrange = range( sapply(1:L, function(i) err[[i]]$abs$day), na.rm=TRUE )
  41     matplot(sapply( seq_len(L), function(i) err[[i]]$abs$day ), type="l",
  42         xlab="Time (hours)", ylab="Mean |y - y_hat|", ylim=yrange, col=cols, lwd=2, lty=1)
  43
  44     agg_curves <- sapply( seq_len(L), function(i) {
  45         curve <- err[[i]]$abs$indices
  46         delta <- if (agg=="day") 1 else if (agg=="week") 7 else if (agg=="month") 30
  47         vapply( seq(1,length(curve),delta), function(i) {
  48             mean(curve[i:(i+delta-1)], na.rm=TRUE)
  49         }, vector("double",1), USE.NAMES=FALSE )
  50     })
  51     yrange = range(agg_curves, na.rm=TRUE)
  52     matplot(agg_curves, type="l", xlab=paste("Time (",agg,"s)", sep=""),
  53         ylab="Mean |y - y_hat|", ylim=yrange, col=cols, lwd=2, lty=1)
  54
  55     yrange = range( sapply(1:L, function(i) err[[i]]$MAPE$day), na.rm=TRUE )
  56     matplot(sapply( seq_len(L), function(i) err[[i]]$MAPE$day ), type="l",
  57         xlab="Time (hours)", ylab="Mean MAPE", ylim=yrange, col=cols, lwd=2, lty=1)
  58
  59     agg_curves <- sapply( seq_len(L), function(i) {
  60         curve <- err[[i]]$MAPE$indices
  61         delta <- if (agg=="day") 1 else if (agg=="week") 7 else if (agg=="month") 30
  62         vapply( seq(1,length(curve),delta), function(i) {
  63             mean(curve[i:(i+delta-1)], na.rm=TRUE)
  64         }, vector("double",1), USE.NAMES=FALSE )
  65     })
  66     yrange = range(agg_curves, na.rm=TRUE)
  67     matplot(agg_curves, type="l", xlab=paste("Time (",agg,"s)", sep=""),
  68         ylab="Mean MAPE", ylim=yrange, col=cols, lwd=2, lty=1)
  69 }
  70
  71 #' Plot measured / predicted
  72 #'
  73 #' Plot measured curve (in black) and predicted curve (in blue).
  74 #'
  75 #' @inheritParams computeError
  76 #' @param index Index in forecasts (integer or date)
  77 #'
  78 #' @export
  79 plotPredReal <- function(data, pred, index)
  80 {
  81     prediction = pred$getForecast(index)
  82     measure = data$getSerie( pred$getIndexInData(index) )[1:length(pred$getForecast(1))]
  83
  84     # Remove the common part, where prediction == measure
  85     dot_mark <- ifelse(prediction[1]==measure[1],
  86         which.max(seq_along(prediction)[prediction==measure]), 0)
  87     prediction = prediction[(dot_mark+1):length(prediction)]
  88     measure = measure[(dot_mark+1):length(measure)]
  89
  90     yrange = range(measure, prediction)
  91     par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=3)
  92     plot(measure, type="l", ylim=yrange, xlab="Time (hours)", ylab="PM10")
  93     par(new=TRUE)
  94     plot(prediction, type="l", col="#0000FF", ylim=yrange, xlab="", ylab="")
  95 }
  96
  97 #' Plot similarities
  98 #'
  99 #' Plot histogram of similarities (weights), for 'Neighbors' method.
 100 #'
 101 #' @inheritParams computeError
 102 #' @param index Index in forecasts (integer or date)
 103 #'
 104 #' @export
 105 plotSimils <- function(pred, index)
 106 {
 107     weights = pred$getParams(index)$weights
 108     if (is.null(weights))
 109         stop("plotSimils only works on 'Neighbors' forecasts")
 110     par(mfrow=c(1,2), mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
 111     small_weights = weights[ weights < 1/length(weights) ]
 112     large_weights = weights[ weights >= 1/length(weights) ]
 113     hist(small_weights, nclass=25, main="", xlab="Weight < 1/N", ylab="Count")
 114     hist(large_weights, nclass=25, main="", xlab="Weight >= 1/N", ylab="Count")
 115 }
 116
 117 #' Functional boxplot
 118 #'
 119 #' Draw the functional boxplot on the left, and bivariate plot on the right.
 120 #'
 121 #' @inheritParams computeError
 122 #' @inheritParams plotCurves
 123 #'
 124 #' @export
 125 plotFbox <- function(data, indices=seq_len(data$getSize()))
 126 {
 127     if (!requireNamespace("rainbow", quietly=TRUE))
 128         stop("Functional boxplot requires the rainbow package")
 129
 130     series_matrix = data$getSeries(indices)
 131     # Remove series with NAs
 132     no_NAs_indices = sapply( 1:ncol(series_matrix),
 133         function(i) all(!is.na(series_matrix[,i])) )
 134     series_matrix = series_matrix[,no_NAs_indices]
 135
 136     series_fds = rainbow::fds(seq_len(nrow(series_matrix)), series_matrix)
 137     par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
 138     rainbow::fboxplot(series_fds, "functional", "hdr", xlab="Time (hours)", ylab="PM10",
 139         plotlegend=FALSE, lwd=2)
 140     rainbow::fboxplot(series_fds, "bivariate", "hdr", plotlegend=FALSE)
 141 }
 142
 143 #' Compute filaments
 144 #'
 145 #' Obtain similar days in the past, and (optionally) plot them -- as black as distances
 146 #' are small.
 147 #'
 148 #' @inheritParams computeError
 149 #' @param index Index in forecast (integer or date)
 150 #' @param limit Number of neighbors to consider
 151 #' @param plot Should the result be plotted?
 152 #'
 153 #' @return A list with
 154 #' \itemize{
 155 #'   \item index : index of the current serie ('today')
 156 #'   \item neighb_indices : indices of its neighbors
 157 #'   \item colors : colors of neighbors curves (shades of gray)
 158 #' }
 159 #'
 160 #' @export
 161 computeFilaments <- function(data, pred, index, limit=60, plot=TRUE)
 162 {
 163     weights <- pred$getParams(index)$weights
 164     if (is.null(weights) || is.na(pred$getParams(index)$weights[1]))
 165         stop("computeFilaments requires a serie without NAs")
 166
 167     nn <- min(limit, length(weights))
 168     sorted_dists = sort(-log(weights), index.return=TRUE)
 169     # Compute colors for each neighbor (from darkest to lightest), if weights differ
 170     if ( any( weights != weights[1] ) )
 171     {
 172         min_dist = min(sorted_dists$x[1:nn])
 173         max_dist = max(sorted_dists$x[1:nn])
 174         color_values = floor(19.5*(sorted_dists$x[1:nn]-min_dist)/(max_dist-min_dist)) + 1
 175         colors = gray.colors(20,0.1,0.9)[color_values] #TODO: 20 == magic number
 176     }
 177     else
 178         colors <- rep(colors()[17], length(weights))
 179
 180     if (plot)
 181     {
 182         # Complete series with (past and present) tomorrows
 183         ref_serie = c( data$getCenteredSerie( pred$getIndexInData(index)-1 ),
 184             data$getCenteredSerie( pred$getIndexInData(index) ) )
 185         centered_series = rbind(
 186             data$getCenteredSeries( pred$getParams(index)$indices-1 ),
 187             data$getCenteredSeries( pred$getParams(index)$indices ) )
 188         yrange = range( ref_serie,
 189             quantile(centered_series, probs=c(0.025,0.975), na.rm=TRUE) )
 190         par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=2)
 191         for (i in nn:1)
 192         {
 193             plot(centered_series[,sorted_dists$ix[i]], ylim=yrange, type="l", col=colors[i],
 194                 xlab=ifelse(i==1,"Time (hours)",""), ylab=ifelse(i==1,"Centered PM10",""))
 195             par(new=TRUE)
 196         }
 197         # Also plot ref curve, in red
 198         plot(ref_serie, ylim=yrange, type="l", col="#FF0000", xlab="", ylab="")
 199         dot_mark <- 0.5 + which.max( pred$getForecast(1) ==
 200             data$getSerie( pred$getIndexInData(1) )[1:length(pred$getForecast(1))] )
 201         abline(v=24+dot_mark, lty=2, col=colors()[56], lwd=1)
 202     }
 203
 204     list(
 205         "index"=pred$getIndexInData(index),
 206         "neighb_indices"=pred$getParams(index)$indices[sorted_dists$ix[1:nn]],
 207         "colors"=colors)
 208 }
 209
 210 #' Functional boxplot on filaments
 211 #'
 212 #' Draw the functional boxplot on filaments obtained by \code{computeFilaments()}.
 213 #'
 214 #' @inheritParams computeError
 215 #' @param fil Output of \code{computeFilaments}
 216 #' @param predict_from First predicted time step
 217 #'
 218 #' @export
 219 plotFilamentsBox = function(data, fil, predict_from)
 220 {
 221     if (!requireNamespace("rainbow", quietly=TRUE))
 222         stop("Functional boxplot requires the rainbow package")
 223
 224     series_matrix = rbind(
 225         data$getSeries(fil$neighb_indices-1), data$getSeries(fil$neighb_indices) )
 226     series_fds = rainbow::fds(seq_len(nrow(series_matrix)), series_matrix)
 227
 228     par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
 229     rainbow::fboxplot(series_fds, "functional", "hdr", xlab="Time (hours)", ylab="PM10",
 230         plotlegend=FALSE, lwd=2)
 231
 232     # "Magic": http://stackoverflow.com/questions/13842560/get-xlim-from-a-plot-in-r
 233     usr <- par("usr")
 234     yr <- (usr[4] - usr[3]) / 27
 235     par(new=TRUE)
 236     plot(c(data$getSerie(fil$index-1),data$getSerie(fil$index)), type="l", lwd=2, lty=2,
 237         ylim=c(usr[3] + yr, usr[4] - yr), xlab="", ylab="")
 238     abline(v=24+predict_from-0.5, lty=2, col=colors()[56])
 239 }
 240
 241 #' Plot relative conditional variability / absolute variability
 242 #'
 243 #' Draw the relative conditional variability / absolute variability based on filaments
 244 #' obtained by \code{computeFilaments()}.
 245 #'
 246 #' @inheritParams computeError
 247 #' @inheritParams plotFilamentsBox
 248 #'
 249 #' @export
 250 plotRelVar = function(data, fil, predict_from)
 251 {
 252     ref_var = c( apply(data$getSeries(fil$neighb_indices-1),1,sd),
 253         apply(data$getSeries(fil$neighb_indices),1,sd) )
 254     tdays = .getNoNA2(data, 2, fil$index)
 255     global_var = c(
 256         apply(data$getSeries(tdays-1),1,sd),
 257         apply(data$getSeries(tdays),1,sd) )
 258
 259     yrange = range(ref_var, global_var)
 260     par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
 261     plot(ref_var, type="l", col=1, lwd=3, ylim=yrange,
 262         xlab="Time (hours)", ylab="Standard deviation")
 263     par(new=TRUE)
 264     plot(global_var, type="l", col=2, lwd=3, ylim=yrange, xlab="", ylab="")
 265     abline(v=24+predict_from-0.5, lty=2, col=colors()[56])
 266 }