ad0ed4ea7e5ac0999d270a10c1dd8ea61d0b7753
[talweg.git] / 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 for (i in seq_along(indices))
15 {
16 plot(series[,i], type="l", ylim=yrange,
17 xlab=ifelse(i==1,"Time (hours)",""), ylab=ifelse(i==1,"PM10",""))
18 if (i < length(indices))
19 par(new=TRUE)
20 }
21 }
22
23 #' Plot error
24 #'
25 #' Draw error graphs, potentially from several runs of \code{computeForecast()}.
26 #'
27 #' @param err Error as returned by \code{computeError()}
28 #' @param cols Colors for each error (default: 1,2,3,...)
29 #'
30 #' @seealso \code{\link{plotCurves}}, \code{\link{plotPredReal}},
31 #' \code{\link{plotSimils}}, \code{\link{plotFbox}}, \code{\link{computeFilaments}},
32 #' \code{\link{plotFilamentsBox}}, \code{\link{plotRelVar}}
33 #'
34 #' @export
35 plotError <- function(err, cols=seq_along(err))
36 {
37 if (!is.null(err$abs))
38 err = list(err)
39 par(mfrow=c(2,2), mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=2)
40 L = length(err)
41 yrange = range( sapply(1:L, function(i) ( err[[i]]$abs$day ) ), na.rm=TRUE )
42 for (i in seq_len(L))
43 {
44 plot(err[[i]]$abs$day, type="l", xlab=ifelse(i==1,"Time (hours)",""),
45 ylab=ifelse(i==1,"Mean |y - y_hat|",""), ylim=yrange, col=cols[i])
46 if (i < L)
47 par(new=TRUE)
48 }
49 yrange = range( sapply(1:L, function(i) ( err[[i]]$abs$indices ) ), na.rm=TRUE )
50 for (i in seq_len(L))
51 {
52 plot(err[[i]]$abs$indices, type="l", xlab=ifelse(i==1,"Time (days)",""),
53 ylab=ifelse(i==1,"Mean |y - y_hat|",""), ylim=yrange, col=cols[i])
54 if (i < L)
55 par(new=TRUE)
56 }
57 yrange = range( sapply(1:L, function(i) ( err[[i]]$MAPE$day ) ), na.rm=TRUE )
58 for (i in seq_len(L))
59 {
60 plot(err[[i]]$MAPE$day, type="l", xlab=ifelse(i==1,"Time (hours)",""),
61 ylab=ifelse(i==1,"Mean MAPE",""), ylim=yrange, col=cols[i])
62 if (i < L)
63 par(new=TRUE)
64 }
65 yrange = range( sapply(1:L, function(i) ( err[[i]]$MAPE$indices ) ), na.rm=TRUE )
66 for (i in seq_len(L))
67 {
68 plot(err[[i]]$MAPE$indices, type="l", xlab=ifelse(i==1,"Time (days)",""),
69 ylab=ifelse(i==1,"Mean MAPE",""), ylim=yrange, col=cols[i])
70 if (i < L)
71 par(new=TRUE)
72 }
73 }
74
75 #' Plot measured / predicted
76 #'
77 #' Plot measured curve (in black) and predicted curve (in blue).
78 #'
79 #' @inheritParams computeError
80 #' @param index Index in forecasts (integer or date)
81 #'
82 #' @export
83 plotPredReal <- function(data, pred, index)
84 {
85 prediction = pred$getForecast(index)
86 measure = data$getSerie( pred$getIndexInData(index) )[length(prediction)]
87 yrange = range(measure, prediction)
88 par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=3)
89 plot(measure, type="l", ylim=yrange, xlab="Time (hours)", ylab="PM10")
90 par(new=TRUE)
91 plot(prediction, type="l", col="#0000FF", ylim=yrange, xlab="", ylab="")
92 }
93
94 #' Plot similarities
95 #'
96 #' Plot histogram of similarities (weights), for 'Neighbors' method.
97 #'
98 #' @inheritParams computeError
99 #' @param index Index in forecasts (integer or date)
100 #'
101 #' @export
102 plotSimils <- function(pred, index)
103 {
104 weights = pred$getParams(index)$weights
105 if (is.null(weights))
106 stop("plotSimils only works on 'Neighbors' forecasts")
107 par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
108 hist(pred$getParams(index)$weights, nclass=20, main="", xlab="Weight", ylab="Count")
109 }
110
111 #' Functional boxplot
112 #'
113 #' Draw the functional boxplot on the left, and bivariate plot on the right.
114 #'
115 #' @inheritParams computeError
116 #' @inheritParams plotCurves
117 #'
118 #' @export
119 plotFbox <- function(data, indices=seq_len(data$getSize()))
120 {
121 if (!requireNamespace("rainbow", quietly=TRUE))
122 stop("Functional boxplot requires the rainbow package")
123
124 series_matrix = data$getSeries(indices)
125 # Remove series with NAs
126 no_NAs_indices = sapply( 1:ncol(series_matrix),
127 function(i) all(!is.na(series_matrix[,i])) )
128 series_matrix = series_matrix[,no_NAs_indices]
129
130 series_fds = rainbow::fds(seq_len(nrow(series_matrix)), series_matrix)
131 par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
132 rainbow::fboxplot(series_fds, "functional", "hdr", xlab="Time (hours)", ylab="PM10",
133 plotlegend=FALSE, lwd=2)
134 rainbow::fboxplot(series_fds, "bivariate", "hdr", plotlegend=FALSE)
135 }
136
137 #' Compute filaments
138 #'
139 #' Obtain similar days in the past, and (optionally) plot them -- as black as distances
140 #' are small.
141 #'
142 #' @inheritParams computeError
143 #' @param index Index in forecast (integer or date)
144 #' @param limit Number of neighbors to consider
145 #' @param plot Should the result be plotted?
146 #' @param predict_from First prediction instant
147 #'
148 #' @return A list with
149 #' \itemize{
150 #' \item index : index of the current serie ('today')
151 #' \item neighb_indices : indices of its neighbors
152 #' \item colors : colors of neighbors curves (shades of gray)
153 #' }
154 #'
155 #' @export
156 computeFilaments <- function(data, pred, index, predict_from, limit=60, plot=TRUE)
157 {
158 if (is.null(pred$getParams(index)$weights) || is.na(pred$getParams(index)$weights[1]))
159 stop("computeFilaments requires a serie without NAs")
160
161 # Compute colors for each neighbor (from darkest to lightest)
162 sorted_dists = sort(-log(pred$getParams(index)$weights), index.return=TRUE)
163 nn = min(limit, length(sorted_dists$x))
164 min_dist = min(sorted_dists$x[1:nn])
165 max_dist = max(sorted_dists$x[1:nn])
166 color_values = floor(19.5*(sorted_dists$x[1:nn]-min_dist)/(max_dist-min_dist)) + 1
167 colors = gray.colors(20,0.1,0.9)[color_values] #TODO: 20 == magic number
168
169 if (plot)
170 {
171 # Complete series with (past and present) tomorrows
172 ref_serie = c( data$getCenteredSerie( pred$getIndexInData(index)-1 ),
173 data$getCenteredSerie( pred$getIndexInData(index) ) )
174 centered_series = rbind(
175 data$getCenteredSeries( pred$getParams(index)$indices ),
176 data$getCenteredSeries( pred$getParams(index)$indices+1 ) )
177 yrange = range( ref_serie,
178 quantile(centered_series, probs=c(0.025,0.975), na.rm=TRUE) )
179 par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=2)
180 for (i in nn:1)
181 {
182 plot(centered_series[,sorted_dists$ix[i]], ylim=yrange, type="l", col=colors[i],
183 xlab=ifelse(i==1,"Time (hours)",""), ylab=ifelse(i==1,"Centered PM10",""))
184 par(new=TRUE)
185 }
186 # Also plot ref curve, in red
187 plot(ref_serie, ylim=yrange, type="l", col="#FF0000", xlab="", ylab="")
188 abline(v=24+predict_from-0.5, lty=2, col=colors()[56], lwd=1)
189 }
190
191 list(
192 "index"=pred$getIndexInData(index),
193 "neighb_indices"=pred$getParams(index)$indices[sorted_dists$ix[1:nn]],
194 "colors"=colors)
195 }
196
197 #' Functional boxplot on filaments
198 #'
199 #' Draw the functional boxplot on filaments obtained by \code{computeFilaments()}.
200 #'
201 #' @inheritParams computeError
202 #' @param fil Output of \code{computeFilaments}
203 #'
204 #' @export
205 plotFilamentsBox = function(data, fil, predict_from)
206 {
207 if (!requireNamespace("rainbow", quietly=TRUE))
208 stop("Functional boxplot requires the rainbow package")
209
210 series_matrix = rbind(
211 data$getSeries(fil$neighb_indices), data$getSeries(fil$neighb_indices+1) )
212 series_fds = rainbow::fds(seq_len(nrow(series_matrix)), series_matrix)
213 par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
214 rainbow::fboxplot(series_fds, "functional", "hdr", xlab="Time (hours)", ylab="PM10",
215 plotlegend=FALSE, lwd=2)
216
217 # "Magic": http://stackoverflow.com/questions/13842560/get-xlim-from-a-plot-in-r
218 usr <- par("usr")
219 yr <- (usr[4] - usr[3]) / 27
220 par(new=TRUE)
221 plot(c(data$getSerie(fil$index),data$getSerie(fil$index+1)), type="l", lwd=2, lty=2,
222 ylim=c(usr[3] + yr, usr[4] - yr), xlab="", ylab="")
223 abline(v=24+predict_from-0.5, lty=2, col=colors()[56])
224 }
225
226 #' Plot relative conditional variability / absolute variability
227 #'
228 #' Draw the relative conditional variability / absolute variability based on filaments
229 #' obtained by \code{computeFilaments()}.
230 #'
231 #' @inheritParams computeError
232 #' @inheritParams plotFilamentsBox
233 #'
234 #' @export
235 plotRelVar = function(data, fil, predict_from)
236 {
237 ref_var = c( apply(data$getSeries(fil$neighb_indices),1,sd),
238 apply(data$getSeries(fil$neighb_indices+1),1,sd) )
239 tdays = .getNoNA2(data, 2, fil$index)
240 global_var = c(
241 apply(data$getSeries(tdays-1),1,sd),
242 apply(data$getSeries(tdays),1,sd) )
243
244 yrange = range(ref_var, global_var)
245 par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
246 plot(ref_var, type="l", col=1, lwd=3, ylim=yrange,
247 xlab="Time (hours)", ylab="Standard deviation")
248 par(new=TRUE)
249 plot(global_var, type="l", col=2, lwd=3, ylim=yrange, xlab="", ylab="")
250 abline(v=24+predict_from-0.5, lty=2, col=colors()[56])
251 }