MLB runs per game per team

gistfile1.r

# THE HISTORICAL RECORD - RUNS PER GAME BY TEAM
#
# public Gist of this code can be found at
# https://gist.github.com/MonkmanMH/4720641
#
# data source: Lahman Database
# http://www.seanlahman.com/baseball-archive/statistics/
# 2012 version (1871-2012)
# data table "Teams.csv"
#
# note: requires that code found at
# https://gist.github.com/4702915
# has already been run, creating the data frame "Teams.merge"
#
# select the team you want from the franchID variable in the Teams.merge data frame
# note the use of double "=" !
Team1 <- as.data.frame(subset (Teams.merge, franchID == "SEA"))
#
# then pull first and last years of the analysis and team name fields
firstyear <- Team1$yearID[1]
lastyear <- tail(Team1$yearID, 1)
#
# either one of these lines of code work to pull the team name
team.name <- Team1$name[1]
team.name <- tail(Team1$name, 1)
# now put the team name, years, and the rest of the title text into a single string
title.text <- paste(team.name, firstyear, "-", lastyear, 
   ": Runs scored relative to the league average")
#
# BASIC PLOT
# index
plot(Team1$R_index ~ Team1$yearID,  
  main=title.text,
  type = "l")
#
#
# TREND LINE -- using loess modeling
# references: 
#  http://princeofslides.blogspot.ca/2011/05/sab-r-metrics-basics-of-loess.html 
#  http://research.stowers-institute.org/efg/R/Statistics/loess.htm
#
# create new object Runs.LO for loess model
RunIndex.LO <- loess(Team1$R_index ~ Team1$yearID)  			
RunIndex.LO.predict <- predict(RunIndex.LO)
#
# plot the data, add loess curve
ylim <- c(60,140)
plot(Team1$R_index ~ Team1$yearID, 
  ylim = ylim, 
  main = title.text, 
  xlab = "year", ylab = "index (league average = 100)")
# chart tidying
  grid()
# loess predicted value line
  lines(Team1$yearID, RunIndex.LO.predict, 
    lty="solid", col="red", lwd=2)	
#
# MODEL VERSION 2 (with more nuance/zig-zag to the trend line)
# create new object RunIndex.LO for loess model, span=0.25 and span=0.5
RunIndex.LO.25 <- loess(Team1$R_index ~ Team1$yearID, span=0.25)				
RunIndex.LO.25.predict <- predict(RunIndex.LO.25)
#
RunIndex.LO.5 <- loess(Team1$R_index ~ Team1$yearID, span=0.5)				
RunIndex.LO.5.predict <- predict(RunIndex.LO.5)
#
# plot the data, add loess curves
ylim <- c(60,140)
plot(Team1$R_index ~ Team1$yearID, 
  ylim = ylim, 
  main = title.text, 
  xlab = "Year", ylab = "Index (league average = 100)")
# loess predicted value line
  lines(Team1$yearID, RunIndex.LO.predict, lty="solid", col="black", lwd=1.5)
  lines(Team1$yearID, RunIndex.LO.25.predict, lty="dashed", col="red", lwd=2)
  lines(Team1$yearID, RunIndex.LO.5.predict, lty="dotdash", col="blue", lwd=1.5)
# chart legend and grid lines
  legend(firstyear+2, 140, 
    c("default", "span=0.25", "span=0.50"), 
    lty=c("solid", "dashed", "dotdash"), 
    col=c("black", "red", "blue"), 
    lwd=c(1.5, 2, 1.5))
  grid()
# straight line at 100
  abline(h = 100, lty="dotdash") 
#
#