R Notebook

library(reshape2)
library(ggplot2)
library(pastecs)
library(psych)
library(mice)
library(VIM)
library(corrplot)

Pull the data

moneyballdata <- read.csv("moneyball-training-data.csv", header = TRUE)
head(moneyballdata)

Summarize

dim(moneyballdata)

[1] 2276   17

Whittle down the data set to wins, hits, doubles, triples

head(mike.data.set)

     wins hits doubles triples
[1,]   39 1445     194      39
[2,]   70 1339     219      22
[3,]   86 1377     232      35
[4,]   70 1387     209      38
[5,]   82 1297     186      27
[6,]   75 1279     200      36

Get details on wins variable

typeof(wins)

[1] "integer"

summary(wins)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
      0      71      82      81      92     146 

boxplot(wins, col = "orange")

hist(wins, col = "orange")

Get details on hits variable

typeof(hits)

[1] "integer"

summary(hits)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
    891    1383    1454    1469    1537    2554 

boxplot(hits, col = "blue")

hist(hits, col = "blue")

Get details on doubles variable

typeof(doubles)

[1] "integer"

summary(doubles)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
     69     208     238     241     273     458 

boxplot(doubles, col = "green")

hist(doubles, col = "green")

Get details on triples variable

typeof(triples)

[1] "integer"

summary(triples)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
      0      34      47      55      72     223 

boxplot(triples, col = "turquoise")

hist(triples, col = "turquoise")

Obtain a more detailed summary of variables

stat.desc(mike.data.set)

Find correlations between pairs

pairs.panels(mike.data.set)

Check for missing values

md.pattern(mike.data.set)

     wins hits doubles triples  
[1,]    1    1       1       1 0
[2,]    0    0       0       0 0

aggr_plot <- aggr(mike.data.set, 
                col=c('navyblue','red'), 
                numbers=TRUE, 
                sortVars=TRUE, 
                labels=names(mike.data.set), 
                cex.axis=.7, 
                gap=3, 
                ylab=c("Histogram of missing data","Pattern"))

 Variables sorted by number of missings: 
 Variable Count
     wins     0
     hits     0
  doubles     0
  triples     0

Produce a scatterplot

plot(hits, wins, main="Scatterplot wins vs. hits")

Check the correlation between wins and hits

cor(wins, hits)

[1] 0.39

Create an example simple linear regression model. I used wins based on hits because it had the highest correlation.

mod <- lm(wins ~ hits)
summary(mod)

Call:
lm(formula = wins ~ hits)

Residuals:
   Min     1Q Median     3Q    Max 
-71.77  -8.76   0.86   9.76  46.02 

Coefficients:
            Estimate Std. Error t value             Pr(>|t|)    
(Intercept)  18.5623     3.1075    5.97         0.0000000027 ***
hits          0.0423     0.0021   20.12 < 0.0000000000000002 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 14 on 2274 degrees of freedom
Multiple R-squared:  0.151, Adjusted R-squared:  0.151 
F-statistic:  405 on 1 and 2274 DF,  p-value: <0.0000000000000002

attributes(mod)

$names
 [1] "coefficients"  "residuals"     "effects"       "rank"          "fitted.values" "assign"       
 [7] "qr"            "df.residual"   "xlevels"       "call"          "terms"         "model"        

$class
[1] "lm"

Check the model’s residuals or errors

par(mfrow=c(2,2))
plot(mod)

Fit the line

plot(wins~hits, data=faithful)
abline(mod, col=2, lwd=3)

Multple regression on wins vs doubles + triples

mod2 <- lm(wins ~ doubles + triples)
summary(mod2)

Call:
lm(formula = wins ~ doubles + triples)

Residuals:
   Min     1Q Median     3Q    Max 
-69.90  -9.38   0.60  10.07  51.32 

Coefficients:
            Estimate Std. Error t value            Pr(>|t|)    
(Intercept) 50.31417    1.81465   27.73 <0.0000000000000002 ***
doubles      0.10365    0.00668   15.51 <0.0000000000000002 ***
triples      0.09904    0.01120    8.85 <0.0000000000000002 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 15 on 2273 degrees of freedom
Multiple R-squared:  0.114, Adjusted R-squared:  0.113 
F-statistic:  146 on 2 and 2273 DF,  p-value: <0.0000000000000002