Gaston Sanchez
# download data set into your working directory
url <- "https://raw.githubusercontent.com/ucb-introstat/introstat-spring-2019/master/data/nba_players.csv"
dat <- read.csv(url)The dataset dat is a data.frame (a tabular object) with dimensions
(i.e. number of rows and columns):
dim(dat)## [1] 528 39
In turn, the available variables (in the columns) are:
colnames(dat)## [1] "player" "player_num"
## [3] "position" "height"
## [5] "weight" "birthdate"
## [7] "country" "experience"
## [9] "college" "team_stat_ranking"
## [11] "age" "games"
## [13] "games_started" "minutes_played"
## [15] "field_goals" "field_goal_attempts"
## [17] "field_goal_percent" "points3"
## [19] "points3_attempts" "points3_percent"
## [21] "points2" "points2_attempts"
## [23] "points2_percent" "effective_field_goal_percent"
## [25] "free_throws" "free_throw_attempts"
## [27] "free_throw_percent" "offensive_rebounds"
## [29] "defensive_rebounds" "total_rebounds"
## [31] "assists" "steals"
## [33] "blocks" "turnovers"
## [35] "fouls" "points"
## [37] "team_salary_ranking" "salary"
## [39] "team"
To take a look at a particular variable, you can type the name of the
table, followed by the $ sign, followed by the name of the variable:
dat$player
dat$positionA common first step in examining the values of a variable is to count
the frequencies of such values. This can be easily done in R with the
function table(). For instance, to get the counts of the different
types of positions, run the following command:
# position frequencies
table(dat$position)##
## center point guard power forward shooting guard small forward
## 96 112 112 107 101
# age frequencies
table(dat$age)##
## 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
## 9 19 19 36 51 57 47 45 41 35 35 34 24 17 16 8 18 3 4 5 5
In order to get relative frequencies (i.e. proportions or fractions) you need to divide the frequencies by the total number of observations. You can do this in R in several ways:
# position relative frequencies (option 1)
table(dat$position) / nrow(dat)##
## center point guard power forward shooting guard small forward
## 0.1818182 0.2121212 0.2121212 0.2026515 0.1912879
# position relative frequencies (option 2)
table(dat$position) / length(dat$position)##
## center point guard power forward shooting guard small forward
## 0.1818182 0.2121212 0.2121212 0.2026515 0.1912879
# position relative frequencies (option 3)
prop.table(table(dat$position))##
## center point guard power forward shooting guard small forward
## 0.1818182 0.2121212 0.2121212 0.2026515 0.1912879
To visualize frequencies (or relative frequencies) you can create a
bar-chart with the function barplot(). All you need is to take a
table() and pass it inside barplot():
barplot(table(dat$position))
Another common graph to visualize frequencies is a pie-chart, which can
be produced in R with the function pie(). Like in a bar-chart, you can
pass a table() to pie():
pie(table(dat$position))
Let’s consider the quantitative variable height, obtaining frequencies
and a bar-chart.
table(dat$height)##
## 69 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
## 2 5 12 22 22 37 38 34 52 51 45 69 55 41 29 8 2 4
barplot(table(dat$height), las = 1, border = NA)
The bar-chart reveals some information but there are some issues. Can you guess what are those issues?
Or consider weight. Things can get a bit overwhelming:
table(dat$weight)##
## 161 165 170 172 173 175 179 180 183 185 186 188 189 190 191 192 193 194
## 1 5 4 3 2 14 1 5 3 21 4 1 3 18 3 2 2 4
## 195 196 197 198 199 200 201 202 203 205 206 207 208 209 210 212 213 214
## 17 1 1 3 3 28 2 1 2 22 6 1 2 4 23 3 4 3
## 215 216 217 218 219 220 221 222 223 225 226 227 228 229 230 231 232 233
## 16 1 1 5 1 27 3 3 1 22 3 1 9 1 21 2 3 1
## 234 235 237 238 239 240 241 242 243 244 245 246 248 249 250 251 252 253
## 2 21 4 3 2 28 2 4 1 1 25 2 3 1 24 2 1 2
## 255 256 257 260 263 265 270 273 275 279 280 285 289 290 307
## 11 1 1 12 2 8 9 2 5 1 1 1 1 1 1
barplot(table(dat$weight))
In these cases, it’s better to reorganize the values into groups or categories. For example, we could form groups 160-180 pounds, 180-200, 200-220, etc:
weight2 <- cut(
dat$weight,
breaks = seq(from = 160, to = 300, by = 20))
table(weight2)## weight2
## (160,180] (180,200] (200,220] (220,240] (240,260] (260,280] (280,300]
## 35 114 124 130 93 28 3
and then make a bar-chart of the categorized variable weight2
barplot(table(weight2), las = 1, border = NA)
In addition to using bar-charts to visualize the frequencies of a quantitative variable, we can also use a histogram.
R has the hist() function that allows you to create a basic histogram:
# default histogram
hist(dat$weight)
We can specify some graphical parameters to customize the appearence of the histogram:
hist(dat$weight, las = 1, col = 'gray80')
More interestingly, we can also play with the argument breaks to
change the number of bars:
hist(dat$weight, las = 1, col = 'gray80', breaks = 8)
Here’s another option:
hist(dat$weight, las = 1, col = 'gray80',
breaks = seq(from = 160, to = 310, by = 10))
And one more histogram with more “resolution”:
hist(dat$weight, las = 1, col = 'gray80',
breaks = seq(from = 160, to = 310, by = 5))