# AUTHOR: Rein Otsason
# TEAM: Samuel Buckstein, Natalie Landon-Brace, Rein Otsason
sink("statsresults.txt", split=TRUE, append=FALSE)
# Import the raw data
data <- read.csv("statsdata.csv", header=TRUE, sep=",", na.strings = "x")
# Data with misses replaced with the penalty value
data.penalty <- data
data.penalty[is.na(data.penalty)] <- 150 # Penalty for missing; arbitrary
# Accuracy metric -- simple average of the 5 shots
accuracy <- (data.penalty$Shot1 + data.penalty$Shot2 + data.penalty$Shot3 + data.penalty$Shot4 + data.penalty$Shot5) / 5
data.accuracy <- data.frame(data, accuracy)
# Vectors of time data
times.10 <- data$Time[which(data$Distance == 10)]
times.20 <- data$Time[which(data$Distance == 20)]
times.30 <- data$Time[which(data$Distance == 30)]
# Vectors of accuracy data
accuracy.10 <- data.accuracy$accuracy[which(data.accuracy$Distance == 10)]
accuracy.20 <- data.accuracy$accuracy[which(data.accuracy$Distance == 20)]
accuracy.30 <- data.accuracy$accuracy[which(data.accuracy$Distance == 30)]
## Time data analysis
# Test for normality of time data
print("Shapiro-Wilks results for time data: ")
shapiro.test(times.10)
shapiro.test(times.20)
shapiro.test(times.30)
# Test homogeneity of variances
var.test(times.10, times.20) # Between 10 m and 20 m
var.test(times.10, times.30) # Between 10 m and 30 m
var.test(times.20, times.30) # Between 20 m and 30 m
# Two-sample Student's t-test, assuming equal variance and independence
t.test(times.10, times.20, var.equal=TRUE, paired=TRUE)
t.test(times.10, times.30, var.equal=TRUE, paired=FALSE)
t.test(times.20, times.30, var.equal=TRUE, paired=FALSE)
## Accuracy data analysis
# Test for normality of accuracy data
print("Shapiro-Wilks results for accuracy data:")
shapiro.test(accuracy.10)
shapiro.test(accuracy.20)
shapiro.test(accuracy.30)
# Test homogeneity of variances
print("F-test for homogeneity of variances:")
var.test(accuracy.10, accuracy.20)
var.test(accuracy.10, accuracy.30)
var.test(accuracy.20, accuracy.30)
# Two-sample Student's t-test
t.test(accuracy.10, accuracy.20, var.equal=FALSE, paired=TRUE)
t.test(accuracy.10, accuracy.30, var.equal=FALSE, paired=FALSE)
t.test(accuracy.20, accuracy.30, var.equal=TRUE, paired=FALSE)
## Summary statistics
print("Accuracy, 10 m:")
print(sprintf("Mean: %f", mean(accuracy.10)))
print(sprintf("variance: %f", var(accuracy.10)))
print("Accuracy, 20 m:")
print(sprintf("Mean: %f", mean(accuracy.20)))
print(sprintf("variance: %f", var(accuracy.20)))
print("Accuracy, 30 m:")
print(sprintf("Mean: %f", mean(accuracy.30)))
print(sprintf("variance: %f", var(accuracy.30)))
print("Time, 10 m:")
print(sprintf("Mean: %f", mean(times.10)))
print(sprintf("variance: %f", var(times.10)))
print("Time, 20 m:")
print(sprintf("Mean: %f", mean(times.20)))
print(sprintf("variance: %f", var(times.20)))
print("Time, 30 m:")
print(sprintf("Mean: %f", mean(times.30)))
print(sprintf("variance: %f", var(times.30)))
sink()
png(filename="experience_boxplot.png")
boxplot(accuracy~Experience, data=data.accuracy, main="Accuracy compared with experience", xlab="Experience", ylab="Accuracy (lower is better)")
dev.off()