-
Notifications
You must be signed in to change notification settings - Fork 1
/
01_scrapeATPn1.R
280 lines (215 loc) · 10.7 KB
/
01_scrapeATPn1.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
library(scrapeR)
library(xts)
library(RColorBrewer)
library(gplots)
library(ggplot2)
library(plyr)
library(extrafont)
loadfonts()
loadSavedN1players <- TRUE
file <- "allATP_1-100.Rdata"
n1.file <- "n1players.Rdata"
# atp plot: http://www.tennisfrontier.com/blogs/el-dude/talent-richness-of-the-atp-era/
# atp stats: http://www.tennis-x.com/stats/atprankhist.shtml
# http://stackoverflow.com/questions/6906661/ggplot2-make-missing-value-in-geom-tile-not-blank
### ------------------------------------------------------------------------------------------------------------- ###
### HELPERS
### ------------------------------------------------------------------------------------------------------------- ###
# ggplot2 theme no fuss
ggtheme <- {
theme_bw() +
#eliminates baground, gridlines, and chart border
theme(
plot.background = element_blank()
,panel.grid.major = element_blank()
,panel.grid.minor = element_blank()
,panel.border = element_blank()
,panel.background = element_blank()
)
}
### ------------------------------------------------------------------------------------------------------------- ###
### 1. load the vector of n1 players
### ------------------------------------------------------------------------------------------------------------- ###
if(loadSavedN1players) {
load(n1.file)
} else {
load(file)
### 1. load scraped tables with freaking all atp top 100 ranking data
t <- as.Date(t, "%d.%m.%Y")
t <- t[length(t):1]
# inverse the list
df <- df[length(df):1]
# get only the player-rank values
tb <- lapply(df, function(tb) {
# discard table header
tb <- tb[-1,]
# create a name, ranking data.frame
tb.splt <- strsplit(tb[,1], "\\\r\\\n(\\\t)+")
table <- do.call(rbind, lapply(tb.splt, function(l) {
if(length(l) == 3) data.frame(player = as.character(l[[3]]), rank = as.numeric(l[[1]])) else stop()
}))
if(is.null(table)) {
NULL
}
else if(nrow(table) == 100 && table[,2] == 1:100) {
as.character(table[,1])
} else {
as.character(table[match(1:100, table[,2]),1])
}
})
### 2. Reduce data
# discard empty tables
lg <- sapply(tb, length)
if(any(lg) != 100) {
tb <- tb[lg==100]
t <- t[lg==100]
}
stopifnot(length(t) == length(tb))
tb <- do.call(cbind, tb)
# get only players which ranked n° 1
player1 <- unique(tb[1,])
player1 <- player1[!is.na(player1)]
if(!file.exists(n1.file)) save(player1, file = n1.file)
}
stopifnot(exists("player1"))
### ------------------------------------------------------------------------------------------------------------- ###
### 2. load the ranking history of each n1 player
### ------------------------------------------------------------------------------------------------------------- ###
player <- structure(gsub("(^.*)\\s\\(.*\\)", "\\1", player1), names = gsub("^.*(\\(.{3}\\))", "\\1", player1))
atpurl.prefix <- "http://www.atpworldtour.com/Tennis/Players/Top-Players/"
atpurl.prefix2 <- "http://www.atpworldtour.com/Tennis/Players/"
atpurl.suffix <- ".aspx?t=rh"
#http://www.atpworldtour.com/Tennis/Players/Top-Players/Andy-Roddick.aspx?t=rh
#http://www.atpworldtour.com/Tennis/Players/Ro/A/Andy-Roddick.aspx
#http://www.atpworldtour.com/Tennis/Players/Na/I/Ilie-Nastase.aspx
n1.player <- lapply(player, function(p) {
cat("\n", p)
p.splt <- strsplit(p, ",\\s")
url.player <- paste(atpurl.prefix, p.splt[[1]][2], "-", p.splt[[1]][1], atpurl.suffix, sep="")
html.tb <- try(readHTMLTable(url.player, stringsAsFactors = FALSE), silent = T)
if(class(html.tb) == "try-error") {
# ugly hacks
if( grepl("Newcombe", p)) {
url.player <- paste(atpurl.prefix2, "Ne/J/John-D-Newcombe", atpurl.suffix, sep="")
html.tb <- try(readHTMLTable(url.player, stringsAsFactors = FALSE), silent = T)
} else if (grepl("Courier", p)) {
url.player <- paste(atpurl.prefix2, "Co/J/Jim-S-Courier", atpurl.suffix, sep="")
html.tb <- try(readHTMLTable(url.player, stringsAsFactors = FALSE), silent = T)
} else {
url.player <- paste(atpurl.prefix2, substr(p.splt[[1]][1], 0, 2), "/", substr(p.splt[[1]][2], 0, 1), "/",
p.splt[[1]][2], "-", p.splt[[1]][1], atpurl.suffix, sep="")
html.tb <- try(readHTMLTable(url.player, stringsAsFactors = FALSE), silent = T)
}
}
stopifnot(is.list(html.tb), length(html.tb) == 1)
html.tb <- html.tb[[1]]
html <- getURL(url.player)
html.list <- strsplit(html,"<ul ")[[1]]
idx <- which(grepl("playerBioInfoList", html.list))
dob <- gsub(".*(\\d{2}\\.\\d{2}\\.\\d{4}).*", "\\1", html.list[idx])
list(atp = html.tb, dob = dob)
})
#save(n1.player, file="n1players_ranking.Rdata")
### ------------------------------------------------------------------------------------------------------------- ###
### 3. load the ranking history of each n1 player
### ------------------------------------------------------------------------------------------------------------- ###
atp <- lapply(n1.player, function(l) {
stopifnot(names(l) == c("atp", "dob"))
atp <- l[['atp']][1:2]
atp <- atp[-c(1),]
# discard empty ranking rows and T values
idx <- !(atp[,2] == "")
atp <- atp[idx,]
#remove comma or trailing T
atp[,2] <- gsub(",", "", atp[,2])
atp[,2] <- gsub("T", "", atp[,2])
zoo(as.numeric(atp[,2]), as.Date(atp[,1], format = "%d.%m.%Y"))
})
names(atp) <- player1
# get all the unique dates
dates <- as.Date(unique(unlist(sapply(atp, function(dff) index(dff)))))
dates <- dates[order(dates)]
# build a matrix of atp ranking with row = dates, col = player
n1.ranking <- zoo(matrix(nrow = length(dates), ncol = length(n1.player)), dates)
colnames(n1.ranking) <- player1
for (i in seq_along(atp)) {
idx <- match(index(atp[[i]]), index(n1.ranking))
n1.ranking[idx,i] <- atp[[i]]
}
### ------------------------------------------------------------------------------------------------------------- ###
### 4. Plot heatmap
### ------------------------------------------------------------------------------------------------------------- ###
date.reg <- c(seq(min(dates), max(dates), 7))
# ugly hack to add another date (for connors he was for one week n2, not visible otherwise)
date.reg <- c(date.reg, as.Date("23.08.1977", "%d.%m.%Y"))
date.reg <- date.reg[order(date.reg)]
idx <- findInterval(date.reg, index(n1.ranking), rightmost.closed = FALSE, all.inside = FALSE)
n1.eqranking <- zoo(unclass(n1.ranking)[idx,], date.reg)
ranks <- max(n1.eqranking, na.rm =T)
# ugly hack to fill ranking before 1985
dateLabel <- rep("", length(date.reg))
ii <- which(!duplicated(format(date.reg, format = "%Y")))
dateLabel[ii[-1]] <- format(date.reg[ii[-1]], "%Y")
dateLabel[ii[1]] <- format(date.reg[ii[1]], "%Y-%m-%d")
#png("fig/heatmap_atp.png", width = 1480, height = 1000, res = 200)
pdf("fig/heatmap_atp.pdf", width = 16, height = 9, family="Georgia")
#svg("fig/heatmap_atp.svg", width = 16, height = 9, family="Georgia", antialias= "none")
heatmap.2(t(log10(n1.eqranking)), Colv = NA, Rowv = NA, density.info = "none", dendrogram = "none", keysize = 1.1, margin = c(7,13),
col = rev(colorRampPalette(brewer.pal(9,"Blues"))(255)), scale = "none", na.rm = T, cexCol = 0.9, trace= "none",
lhei = c(0.8, 5), lwid = c(0.9,6), cexRow = 1.1, labCol = dateLabel, rowsep = 1:ncol(n1.eqranking), sepwidth = c(0, 0.01))
dev.off()
# age <- sapply(1:ncol(n1.ranking), function(i) as.numeric(diff(index(n1.ranking)[range(which(!is.na(n1.ranking[,i])))])) / 365.25 )
### ------------------------------------------------------------------------------------------------------------- ###
### 5. Aging curve
### ------------------------------------------------------------------------------------------------------------- ###
dob <- sapply(n1.player, function(l) {
stopifnot(names(l) == c("atp", "dob"))
as.Date(l[['dob']], format = "%d.%m.%Y")
})
# very ugly hack for
idxna <- sapply(dob, nchar) >=1
if(any(is.na(idxna)) && which(is.na(idxna))==21) {
dob[[21]] <- as.Date("12.02.1980)", "%d.%M.%Y")
}
age.day <- do.call(cbind, lapply(dob, function(a) as.numeric(index(n1.ranking) - a)))
age <- age.day / 365.25
stopifnot(dim(age)==dim(n1.ranking))
n1.ar <- do.call(rbind, lapply(1:ncol(age), function(a) data.frame(age = age[,a], ranking = n1.ranking[,a], player = player[a])))
n1.ar <- n1.ar[which(!is.na(n1.ar$ranking)),]
# get the youngest and oldest age n1
n1.ar$color <- NA
#p.young <- n1.ar[which(n1.ar$age == min(n1.ar[n1.ar$ranking == 1,'age']) & n1.ar$ranking == 1), 'player']
iiidx <- which(n1.ar$age == max(n1.ar[n1.ar$ranking == 1,'age']) & n1.ar$ranking == 1)
p.old <- as.character(n1.ar[iiidx, 'player'])
n1.ar2 <- n1.ar[!n1.ar$player %in% levels(n1.ar$player)[1:4],]
n1.ar2$rage <- round(n1.ar2$age)
n1.ar3 <- ddply(n1.ar2, .(player,rage, color), summarize, yranking = mean(ranking, na.rm = T))
n1.ar3[n1.ar3$player == p.old,'color'] <- 'oldest'
#n1.ar3[n1.ar3$player == p.young,'color'] <- 'youngest'
# find the players still active
ip <- which(!is.na(n1.eqranking[nrow(n1.eqranking),]))
playerAge <- data.frame(player = c(player[ip], p.old ), age = c(sapply(dob[ip], function(d) Sys.Date()- d) / 365.25, n1.ar[iiidx, 'age']))
png("fig/agingRanking_%02d.png", width = 1480, height = 1000, res = 180)
ggplot(data = n1.ar3, aes(rage, yranking)) + geom_smooth(data = n1.ar2, aes(age, ranking), se = T, size = 1) + xlim(15, 35) + ylim (-30, 700) +
geom_line(aes(rage, yranking, group = player), colour = "grey", n = 100, se = F, size = 0.1) + ggtheme
ggplot(data = n1.ar3, aes(rage, yranking)) + geom_smooth(data = n1.ar2, aes(age, ranking), se = T, size = 1) + xlim(15, 35) + ylim (-30, 1000) +
geom_smooth(aes(rage, yranking, group = player, colour = color), colour = "grey", n = 100, se = F, size = 0.15, span = 0.2) + ggtheme +
geom_smooth(data =n1.ar3[!is.na(n1.ar3$color),], aes (rage, yranking, group = player, colour = color), n = 100, se = F, size = 0.2, span = 0.2) +
xlab("age") + ylab("yearly average ATP ranking")
ggplot(data = n1.ar3, aes(rage, yranking)) + geom_smooth(data = n1.ar2, aes(age, ranking), se = T, size = 1) + xlim(15, 35) + ylim (-30, 1000) +
geom_smooth(aes(rage, yranking, group = player, colour = color), colour = "grey", n = 100, se = F, size = 0.15, span = 0.2) + ggtheme +
geom_smooth(data =n1.ar3[!is.na(n1.ar3$color),], aes (rage, yranking, group = player, colour = color), n = 100, se = F, size = 0.2, span = 0.2)+
geom_vline(data = playerAge, aes(xintercept = age, colour = player), alpha = 0.5)
i <- 1
while(i <= nlevels(n1.ar3$player)) {
if (i %% 5 == 0) {
p <- levels(n1.ar3$player)[(i-4):i]
#browser()
gp <- ggplot(data = n1.ar3[n1.ar3$player %in% p,], aes(rage, yranking)) + geom_smooth(data = n1.ar2, aes(age, ranking), se = T, size = 1) +
xlim(15, 35) + ylim (-30, 700) + geom_smooth(aes(rage, yranking, group = player, group = player, colour = player),
n = 100, se = F, size = 0.3, span = 0.2) + ggtheme
print(gp)
}
i <- i +1
}
dev.off()