Week-3: Data Dive
library(tidyverse)## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.4 ✔ readr 2.1.4
## ✔ forcats 1.0.0 ✔ stringr 1.5.1
## ✔ ggplot2 3.4.4 ✔ tibble 3.2.1
## ✔ lubridate 1.9.3 ✔ tidyr 1.3.0
## ✔ purrr 1.0.2
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorsData_set <- "/Users/ba/Documents/IUPUI/Masters/First Sem/Statistics/Dataset/PitchingPost.csv"
Pitching_Data <- read.csv(Data_set)QUESTION-1: Let’s find out who is the best pitcher in World Series (WS) playoff game by looking at Run Allowed Statistic
WS_Pitching_RA<-
Pitching_Data |>
filter(round=='WS') |>
group_by(playerID) |>
summarise(Runs_Allowed = round(mean(R),2))
head(WS_Pitching_RA,10)## # A tibble: 10 × 2
## playerID Runs_Allowed
## <chr> <dbl>
## 1 abreubr01 0
## 2 aceveal01 0
## 3 adamsmi03 1
## 4 affelje01 0.25
## 5 albural01 1
## 6 alexasc01 2
## 7 allenco01 0
## 8 alvarjo03 4
## 9 anderbr02 2
## 10 anderia01 0Lets create a bins to group them into three different categories(“Best”, “Average”, “Worst”) based on their performance.
WS_Pitching_RA$Performance = cut(WS_Pitching_RA$Runs_Allowed, breaks = c(-0.1,4,7,Inf), labels = c('Best','Average','Worst'))
head(WS_Pitching_RA,10)## # A tibble: 10 × 3
## playerID Runs_Allowed Performance
## <chr> <dbl> <fct>
## 1 abreubr01 0 Best
## 2 aceveal01 0 Best
## 3 adamsmi03 1 Best
## 4 affelje01 0.25 Best
## 5 albural01 1 Best
## 6 alexasc01 2 Best
## 7 allenco01 0 Best
## 8 alvarjo03 4 Best
## 9 anderbr02 2 Best
## 10 anderia01 0 BestLets see how probable these performance categories are likely to occur
WS_Pithing_RA_Prob<- count(WS_Pitching_RA, Performance)
WS_Probability_RA <- WS_Pithing_RA_Prob |>
mutate(Probability = round(n/sum(n),2))
WS_Probability_RA## # A tibble: 3 × 3
## Performance n Probability
## <fct> <int> <dbl>
## 1 Best 330 0.87
## 2 Average 41 0.11
## 3 Worst 10 0.03Since it is a world series finding a player with best performance has the highest probability
Lets visualize the about stats
WS_Probability_RA |>
ggplot(aes(x = Performance,y=Probability, fill = Performance)) +
geom_bar(stat = 'identity')+
geom_text(aes(label = Probability), size = 4, vjust = -0.5, hjust = 0.5)+
theme_classic()
Now lets add these probabilities to WS_Pitching_RA table.
WS_Pitching_RA_tag <- left_join(WS_Pitching_RA,WS_Probability_RA, by = 'Performance')
WS_Pitching_RA_tag |>
select(playerID, Runs_Allowed,Performance,Probability)## # A tibble: 381 × 4
## playerID Runs_Allowed Performance Probability
## <chr> <dbl> <fct> <dbl>
## 1 abreubr01 0 Best 0.87
## 2 aceveal01 0 Best 0.87
## 3 adamsmi03 1 Best 0.87
## 4 affelje01 0.25 Best 0.87
## 5 albural01 1 Best 0.87
## 6 alexasc01 2 Best 0.87
## 7 allenco01 0 Best 0.87
## 8 alvarjo03 4 Best 0.87
## 9 anderbr02 2 Best 0.87
## 10 anderia01 0 Best 0.87
## # ℹ 371 more rowsLets visualize this in the bar graphs:
WS_Pitching_RA_tag |>
ggplot(aes(x=Performance,y=Runs_Allowed,color=Performance))+
geom_boxplot()+
theme_classic()
With the Runs Allowed statistic determining whether a player is good or bad at the game won’t be that reliable because runs allowed could also be due to the performance of fielders in a game. So let’s move on to another metric to assess the performance.
QUESTION-2: Let’s find out who is the best pitcher in World Series (WS) playoff game by looking at Earned Runs (A run for which the pitcher is held accountable, excluding runs scored due to errors, passed balls, or other defensive mistakes.)
WS_Pitching_ER<-
Pitching_Data |>
filter(round=='WS') |>
group_by(playerID) |>
summarise(Earned_Runs = round(mean(ER),2))
head(WS_Pitching_ER,10)## # A tibble: 10 × 2
## playerID Earned_Runs
## <chr> <dbl>
## 1 abreubr01 0
## 2 aceveal01 0
## 3 adamsmi03 1
## 4 affelje01 0.25
## 5 albural01 1
## 6 alexasc01 2
## 7 allenco01 0
## 8 alvarjo03 4
## 9 anderbr02 2
## 10 anderia01 0Lets create a bins to group them into three different categories(“Best”, “Average”, “Worst”) based on their performance.
WS_Pitching_ER$Performance = cut(WS_Pitching_ER$Earned_Runs, breaks = c(-0.1,4,7,Inf), labels = c('Best','Average','Worst'))
head(WS_Pitching_ER,10)## # A tibble: 10 × 3
## playerID Earned_Runs Performance
## <chr> <dbl> <fct>
## 1 abreubr01 0 Best
## 2 aceveal01 0 Best
## 3 adamsmi03 1 Best
## 4 affelje01 0.25 Best
## 5 albural01 1 Best
## 6 alexasc01 2 Best
## 7 allenco01 0 Best
## 8 alvarjo03 4 Best
## 9 anderbr02 2 Best
## 10 anderia01 0 BestLet’s see how probable these are to occur
WS_Pithing_ER_Prob<- count(WS_Pitching_ER, Performance)
WS_Probability_ER <- WS_Pithing_ER_Prob |>
mutate(Probability = round(n/sum(n),2))
WS_Probability_ER## # A tibble: 3 × 3
## Performance n Probability
## <fct> <int> <dbl>
## 1 Best 336 0.88
## 2 Average 37 0.1
## 3 Worst 8 0.02WS_Probability_ER |>
ggplot(aes(x = Performance,y=Probability, fill = Performance)) +
geom_bar(stat = 'identity')+
geom_text(aes(label = Probability), size = 4, vjust = -0.5, hjust = 0.5)+
theme_classic()
WS_Pitching_ER_tag <- left_join(WS_Pitching_ER,WS_Probability_ER, by = 'Performance')
WS_Pitching_ER_tag |>
select(playerID, Earned_Runs,Performance,Probability)## # A tibble: 381 × 4
## playerID Earned_Runs Performance Probability
## <chr> <dbl> <fct> <dbl>
## 1 abreubr01 0 Best 0.88
## 2 aceveal01 0 Best 0.88
## 3 adamsmi03 1 Best 0.88
## 4 affelje01 0.25 Best 0.88
## 5 albural01 1 Best 0.88
## 6 alexasc01 2 Best 0.88
## 7 allenco01 0 Best 0.88
## 8 alvarjo03 4 Best 0.88
## 9 anderbr02 2 Best 0.88
## 10 anderia01 0 Best 0.88
## # ℹ 371 more rowsNow usually in the traditional world series there are 7 matches and earned runs are the total in all the matches in a series, so with this metric we won’t be able to access the pitchers performance in a specific match of 9 innings.
QUESTION-3: Let’s find out who is the best pitcher in World Series (WS) playoff game by looking at Earned Run Average (Calculated by taking the number of earned runs a pitcher allows and dividing it by the total number of innings pitched, then multiplying the result by 9 (since a standard baseball game consists of 9 innings). This gives us a clear sign on how the performance of each pitcher is in a single match
WS_Pitching_ERA<-
Pitching_Data |>
filter(round=='WS') |>
group_by(playerID) |>
summarise(Earned_Runs = sum(ER),
IPOuts = sum(IPouts))|>
mutate(New_ERA = round((9*Earned_Runs)/(IPOuts),2))|>
mutate(Inning_Played = IPOuts*3)
WS_Pitching_ERA## # A tibble: 381 × 5
## playerID Earned_Runs IPOuts New_ERA Inning_Played
## <chr> <int> <int> <dbl> <dbl>
## 1 abreubr01 0 15 0 45
## 2 aceveal01 0 6 0 18
## 3 adamsmi03 1 6 1.5 18
## 4 affelje01 1 35 0.26 105
## 5 albural01 1 6 1.5 18
## 6 alexasc01 2 4 4.5 12
## 7 allenco01 0 18 0 54
## 8 alvarjo03 4 10 3.6 30
## 9 anderbr02 2 16 1.12 48
## 10 anderia01 0 15 0 45
## # ℹ 371 more rowsLets only consider players who have played more than 63 innings because in a world series of 7 matches we have 63 inning a succesfull pitcher can play
WS_Pitching_ERAA <- WS_Pitching_ERA |>
filter(Inning_Played > 63)
head(WS_Pitching_ERAA,10)## # A tibble: 10 × 5
## playerID Earned_Runs IPOuts New_ERA Inning_Played
## <chr> <int> <int> <dbl> <dbl>
## 1 affelje01 1 35 0.26 105
## 2 arrieja01 3 34 0.79 102
## 3 baezpe01 4 24 1.5 72
## 4 batismi01 0 24 0 72
## 5 bauertr01 5 25 1.8 75
## 6 beckejo02 3 70 0.39 210
## 7 blantjo01 6 36 1.5 108
## 8 buehlwa01 1 39 0.23 117
## 9 buehrma01 4 22 1.64 66
## 10 bumgama01 1 108 0.08 324according to MLB the ERA < 2.5 is considered as good pitcher. so lets create a bins to group them into two different categories(“Best”, “Average”) based on their performance.
WS_Pitching_ERAA$Performance = cut(WS_Pitching_ERAA$New_ERA, breaks = c(-0.1,2.5,Inf), labels = c('Best','Average'))
head(WS_Pitching_ERAA,10)## # A tibble: 10 × 6
## playerID Earned_Runs IPOuts New_ERA Inning_Played Performance
## <chr> <int> <int> <dbl> <dbl> <fct>
## 1 affelje01 1 35 0.26 105 Best
## 2 arrieja01 3 34 0.79 102 Best
## 3 baezpe01 4 24 1.5 72 Best
## 4 batismi01 0 24 0 72 Best
## 5 bauertr01 5 25 1.8 75 Best
## 6 beckejo02 3 70 0.39 210 Best
## 7 blantjo01 6 36 1.5 108 Best
## 8 buehlwa01 1 39 0.23 117 Best
## 9 buehrma01 4 22 1.64 66 Best
## 10 bumgama01 1 108 0.08 324 BestWS_Pitching_ER_Prob<- count(WS_Pitching_ERAA, Performance)
WS_Probability_ER <- WS_Pitching_ER_Prob |>
mutate(Probability = round(n/sum(n),2))
WS_Probability_ER## # A tibble: 2 × 3
## Performance n Probability
## <fct> <int> <dbl>
## 1 Best 95 0.94
## 2 Average 6 0.06WS_Probability_ER |>
ggplot(aes(x = Performance,y=Probability, fill = Performance)) +
geom_bar(stat = 'identity')+
geom_text(aes(label = Probability), size = 4, vjust = -0.5, hjust = 0.5)+
theme_classic()
From the above data we can see pitchers who have really good performance