Did you know that you can navigate the posts by swiping left and right?
Virat Kohli Vs MS Dhoni -- Statistically!
21 Nov 2016
. category:
math
.
Comments
#data science
#kaggle
#cricket
A Statistical Analysis on two of the biggest stars of IPL
matches = read.csv("matches.csv", stringsAsFactors = TRUE)
deliveries = read.csv("deliveries.csv", stringsAsFactors = TRUE)
df = merge(matches, deliveries, by.x = "id", by.y = "match_id")
df$season = as.factor(df$season)
vk = subset(df, df$batsman == "V Kohli")
msd = subset(df, df$batsman == "MS Dhoni")
#Plotting Kohli Vs Dhoni runs by season
vk_season = aggregate(batsman_runs ~ season, data = vk, FUN = sum)
colnames(vk_season) = c("season", "runs_kohli")
msd_season = aggregate(batsman_runs ~ season, data = msd, FUN = sum)
colnames(msd_season) = c("season", "runs_dhoni")
vk_msd_season = merge(vk_season, msd_season)
library(ggplot2)
library(reshape2)
library(RColorBrewer)
vk_msd_season_long = melt(vk_msd_season) #Transforms the data frame to one, with dhoni/kohli as factors
## Using season as id variables
ggplot(vk_msd_season_long, aes(x = season, y = value, fill = variable)) +
geom_bar(stat="identity", position = "dodge") + #dodge -- place bars side-to-side
scale_fill_manual(values = c("red","yellow")) + #scale_fill_manual for barplots
ggtitle("Kohli vs Dhoni -- Runs by Seasons") +
labs(x = "Season", y = "Runs")
#Dismissal Analysis
vk_dismissal = subset(vk, vk$player_dismissed == "V Kohli")[,c("season", "dismissal_kind")]
vk_dismissal_long = melt(vk_dismissal)
## Using season, dismissal_kind as id variables
ggplot(vk_dismissal_long, aes(x = season, y = ..count.. , fill = dismissal_kind)) +
geom_bar(stat="count") +
ggtitle("Kohli -- Dismissals by Seasons") +
labs(x = "Season", y = "Dismissal Kind")+
scale_fill_brewer(palette = "Set2")
msd_dismissal = subset(msd, msd$player_dismissed == "MS Dhoni")[,c("season", "dismissal_kind")]
msd_dismissal_long = melt(msd_dismissal)
## Using season, dismissal_kind as id variables
ggplot(msd_dismissal_long, aes(x = season, y = ..count.. , fill = dismissal_kind)) +
geom_bar(stat="count") +
ggtitle("Dhoni -- Dismissals by Seasons") +
labs(x = "Season", y = "Dismissal Kind")+
scale_fill_brewer(palette = "Set2")
#Dismissal counts -- by seasons
vk_dismissal_count = as.data.frame(table(vk_dismissal$season))
colnames(vk_dismissal_count) = c("season", "vk_dismissal")
msd_dismissal_count = as.data.frame(table(msd_dismissal$season))
colnames(msd_dismissal_count) = c("season", "msd_dismissal")
vk_msd_dismissal = merge(vk_dismissal_count, msd_dismissal_count)
vk_msd_dismissal_long = melt(vk_msd_dismissal)
## Using season as id variables
ggplot(vk_msd_dismissal_long, aes(x = season, y = value, fill = variable)) +
geom_bar(stat="identity", position = "dodge") +
scale_fill_manual(values = c("red","yellow")) +
ggtitle("Kohli vs Dhoni -- Dismissals by Seasons") +
labs(x = "Season", y = "Dismissals")
#Share of runs made
msd_share = as.data.frame(table(msd$total_runs))
#pos = cumsum(msd_share$Freq) - msd_share$Freq/2
ggplot(msd_share, aes(x = factor(1), y = msd_share$Freq, fill = factor(msd_share$Var1))) +
geom_bar(stat = "identity")+
coord_polar(theta = "y")+
ggtitle("Dhoni's Run Share")+
labs(x = "",y = "")+
scale_fill_discrete(guide_legend(title = "Run Color"))
#geom_text(aes(x= factor(1), y=pos, label = factor(msd_share$Freq)), size=3)
vk_share = as.data.frame(table(vk$total_runs))
#pos = cumsum(vk_share$Freq) - vk_share$Freq/2
ggplot(vk_share, aes(x = factor(1), y = vk_share$Freq, fill = factor(vk_share$Var1))) +
geom_bar(stat = "identity")+
coord_polar(theta = "y")+
ggtitle("Kohli's Run Share")+
labs(x="", y="")+
scale_fill_discrete(guide_legend(title = "Run Color"))
#geom_text(aes(x= factor(1), y=pos, label = factor(vk_share$Freq)), size=3)
#Favorite/Worst team
vk$opposition = vk$team1
vk$opposition[vk$opposition == "Royal Challengers Bangalore"] = vk$team2[vk$opposition == "Royal Challengers Bangalore"]
vk_fav_team = aggregate(batsman_runs ~ opposition, data = vk, FUN = sum)
vk_fav_team[with(vk_fav_team, order(-batsman_runs)),]
## opposition batsman_runs
## 1 Chennai Super Kings 706
## 3 Delhi Daredevils 603
## 5 Kings XI Punjab 477
## 12 Sunrisers Hyderabad 439
## 7 Kolkata Knight Riders 386
## 8 Mumbai Indians 365
## 2 Deccan Chargers 306
## 10 Rajasthan Royals 258
## 4 Gujarat Lions 209
## 11 Rising Pune Supergiants 188
## 9 Pune Warriors 128
## 6 Kochi Tuskers Kerala 50
vk_balls_faced = as.data.frame(table(vk$opposition))
colnames(vk_balls_faced) = c("opposition", "balls_faced")
vk_fav_team = merge(vk_fav_team, vk_balls_faced)
vk_fav_team$average_runs_per_ball = vk_fav_team$batsman_runs/vk_fav_team$balls_faced
vk_fav_team[with(vk_fav_team, order(-average_runs_per_ball)),c("opposition", "average_runs_per_ball")]
## opposition average_runs_per_ball
## 4 Gujarat Lions 1.7272727
## 11 Rising Pune Supergiants 1.4687500
## 12 Sunrisers Hyderabad 1.4070513
## 3 Delhi Daredevils 1.3735763
## 2 Deccan Chargers 1.3304348
## 9 Pune Warriors 1.2673267
## 5 Kings XI Punjab 1.2486911
## 1 Chennai Super Kings 1.2385965
## 8 Mumbai Indians 1.2046205
## 7 Kolkata Knight Riders 1.2024922
## 6 Kochi Tuskers Kerala 0.9615385
## 10 Rajasthan Royals 0.9280576
msd$opposition = msd$team1
msd$opposition[msd$opposition == "Chennai Super Kings"] = msd$team2[msd$opposition == "Chennai Super Kings"]
msd$opposition[msd$opposition == "Rising Pune Supergiants"] = msd$team2[msd$opposition == "Rising Pune Supergiants"]
msd_fav_team = aggregate(batsman_runs ~ opposition, data = msd, FUN = sum)
msd_fav_team[with(msd_fav_team, order(-batsman_runs)),]
## opposition batsman_runs
## 10 Royal Challengers Bangalore 559
## 7 Mumbai Indians 512
## 4 Kings XI Punjab 448
## 2 Delhi Daredevils 403
## 6 Kolkata Knight Riders 359
## 1 Deccan Chargers 281
## 9 Rajasthan Royals 249
## 11 Sunrisers Hyderabad 249
## 8 Pune Warriors 135
## 3 Gujarat Lions 52
## 5 Kochi Tuskers Kerala 23
msd_balls_faced = as.data.frame(table(msd$opposition))
colnames(msd_balls_faced) = c("opposition", "balls_faced")
msd_fav_team = merge(msd_fav_team, msd_balls_faced)
msd_fav_team$average_runs_per_ball = msd_fav_team$batsman_runs/msd_fav_team$balls_faced
msd_fav_team[with(msd_fav_team, order(-average_runs_per_ball)),c("opposition", "average_runs_per_ball")]
## opposition average_runs_per_ball
## 3 Gujarat Lions 1.733333
## 4 Kings XI Punjab 1.523810
## 8 Pune Warriors 1.500000
## 11 Sunrisers Hyderabad 1.500000
## 2 Delhi Daredevils 1.404181
## 7 Mumbai Indians 1.358090
## 10 Royal Challengers Bangalore 1.337321
## 6 Kolkata Knight Riders 1.334572
## 5 Kochi Tuskers Kerala 1.210526
## 9 Rajasthan Royals 1.180095
## 1 Deccan Chargers 1.089147
#Favorite/Worst bowler
vk_fav_bowler = aggregate(batsman_runs ~ bowler, data = vk, FUN = sum)
vk_balls_faced_bowler = as.data.frame(table(vk$bowler))
colnames(vk_balls_faced_bowler) = c("bowler", "balls_faced")
vk_fav_bowler = merge(vk_fav_bowler, vk_balls_faced_bowler)
vk_fav_bowler = subset(vk_fav_bowler, vk_fav_bowler$balls_faced >= 20) #Consider bowlers with more than 20 deliveries
vk_fav_bowler$average_against_bowler = vk_fav_bowler$batsman_runs/vk_fav_bowler$balls_faced
head(vk_fav_bowler[with(vk_fav_bowler, order(-average_against_bowler)),c("bowler", "average_against_bowler")])
## bowler average_against_bowler
## 84 KV Sharma 2.121212
## 167 UT Yadav 1.826667
## 138 Sandeep Sharma 1.652174
## 117 PJ Sangwan 1.650000
## 132 RP Singh 1.589744
## 110 NLTC Perera 1.571429
tail(vk_fav_bowler[with(vk_fav_bowler, order(-average_against_bowler)),c("bowler", "average_against_bowler")])
## bowler average_against_bowler
## 16 Ankit Sharma 0.8518519
## 70 JH Kallis 0.8285714
## 100 MM Patel 0.8000000
## 160 S Sreesanth 0.8000000
## 159 SR Watson 0.7307692
## 152 SK Warne 0.5714286
msd_fav_bowler = aggregate(batsman_runs ~ bowler, data = msd, FUN = sum)
msd_balls_faced_bowler = as.data.frame(table(msd$bowler))
colnames(msd_balls_faced_bowler) = c("bowler", "balls_faced")
msd_fav_bowler = merge(msd_fav_bowler, msd_balls_faced_bowler)
msd_fav_bowler = subset(msd_fav_bowler, msd_fav_bowler$balls_faced >= 20) #Consider bowlers with more than 20 deliveries
msd_fav_bowler$average_against_bowler = msd_fav_bowler$batsman_runs/msd_fav_bowler$balls_faced
head(msd_fav_bowler[with(msd_fav_bowler, order(-average_against_bowler)),c("bowler", "average_against_bowler")])
## bowler average_against_bowler
## 56 IK Pathan 1.904762
## 14 A Nehra 1.900000
## 76 KA Pollard 1.781818
## 47 DW Steyn 1.775862
## 29 CH Gayle 1.760000
## 84 L Balaji 1.760000
tail(msd_fav_bowler[with(msd_fav_bowler, order(-average_against_bowler)),c("bowler", "average_against_bowler")])
## bowler average_against_bowler
## 119 R Bhatia 0.8750000
## 10 A Kumble 0.8518519
## 121 RE van der Merwe 0.8421053
## 40 DL Vettori 0.8214286
## 55 HV Patel 0.8076923
## 155 SP Narine 0.4864865
#Favorite Non-Striker
msd_fav_non_striker = aggregate(batsman_runs ~ non_striker, data = msd, FUN = sum)
head(msd_fav_non_striker[with(msd_fav_non_striker, order(-batsman_runs)),])
## non_striker batsman_runs
## 31 SK Raina 522
## 25 RA Jadeja 349
## 29 S Badrinath 341
## 7 DJ Bravo 291
## 17 ML Hayden 215
## 12 JA Morkel 210
vk_fav_non_striker = aggregate(batsman_runs ~ non_striker, data = vk, FUN = sum)
head(vk_fav_non_striker[with(vk_fav_non_striker, order(-batsman_runs)),])
## non_striker batsman_runs
## 1 AB de Villiers 1012
## 6 CH Gayle 1009
## 19 KL Rahul 277
## 14 JH Kallis 181
## 46 SS Tiwary 154
## 28 MC Henriques 142
#Favorite Venue
msd_fav_venue = aggregate(batsman_runs ~ venue, data = msd, FUN = sum)
head(msd_fav_venue[with(msd_fav_venue, order(-batsman_runs)),])
## venue batsman_runs
## 11 MA Chidambaram Stadium, Chepauk 1144
## 13 M Chinnaswamy Stadium 281
## 6 Eden Gardens 214
## 30 Wankhede Stadium 177
## 9 JSCA International Stadium Complex 138
## 7 Feroz Shah Kotla 127
vk_fav_venue = aggregate(batsman_runs ~ venue, data = vk, FUN = sum)
head(vk_fav_venue[with(vk_fav_venue, order(-batsman_runs)),])
## venue batsman_runs
## 11 M Chinnaswamy Stadium 1763
## 17 Rajiv Gandhi International Stadium, Uppal 298
## 9 MA Chidambaram Stadium, Chepauk 285
## 28 Wankhede Stadium 280
## 5 Feroz Shah Kotla 277
## 4 Eden Gardens 176
#Pacing the innings
vk_over_runs = aggregate(batsman_runs ~ over, data = vk, FUN = sum)
vk_overs_faced = as.data.frame(table(vk$over))
colnames(vk_overs_faced) = c("over", "vk_freq")
vk_over_runs = merge(vk_over_runs, vk_overs_faced)
vk_over_runs$vk_strike_rate = (vk_over_runs$batsman_runs/vk_over_runs$vk_freq)*100
msd_over_runs = aggregate(batsman_runs ~ over, data = msd, FUN = sum)
msd_overs_faced = as.data.frame(table(msd$over))
colnames(msd_overs_faced) = c("over", "msd_freq")
msd_over_runs = merge(msd_over_runs, msd_overs_faced)
msd_over_runs$msd_strike_rate = (msd_over_runs$batsman_runs/msd_over_runs$msd_freq)*100
vk_msd_strike_rate = merge(vk_over_runs[,c("over", "vk_strike_rate")], msd_over_runs[,c("over", "msd_strike_rate")])
vk_msd_strike_rate$over = as.factor(vk_msd_strike_rate$over)
vk_msd_strike_rate_long = melt(vk_msd_strike_rate)
## Using over as id variables
ggplot(vk_msd_strike_rate_long, aes(over, value, group = variable, col = variable)) +
geom_point() + geom_smooth()+
ggtitle("Kohli vs Dhoni -- Strike Rate by Over") +
labs(x = "Over", y = "Strike Rate")
vk_msd_freq = merge(vk_over_runs[,c("over", "vk_freq")], msd_over_runs[,c("over", "msd_freq")])
vk_msd_freq$over = as.factor(vk_msd_freq$over)
vk_msd_freq_long = melt(vk_msd_freq)
## Using over as id variables
ggplot(vk_msd_freq_long, aes(over, value, group = variable, col = variable)) +
geom_point() + geom_smooth()+
ggtitle("Kohli vs Dhoni -- Number of Overs faced") +
labs(x = "Over", y = "Count")
Shubhankar is an awesome person. He's Co-Founder at Houseware, building the next generation of Analytics. In his spare time, he likes to go out on runs!