• Introduction

  • Required Packages
  • This Vignette
  • The functions
  • Pulling Data
  • The Graphs

  knitr::opts_chunk$set(fig.path='Figs/')

  knitr::opts_chunk$set(echo = TRUE, include = TRUE) #install.packages("httr") #install.packages("jsonlite") require(httr) require(jsonlite) require(ggplot2) require(tidyr) require(dplyr)

Introduction

Required Packages

The following packages are required for running this Vignette: -httr -jsonlite -ggplot2 -tidyr -dplyr

This Vignette

This vignette discusses a function that was produced to allow for the relatively easy querying of the NHL records and stats API’s. These API’s have statistics and information, both current and historical, for all the teams that are currently in or were once a part of the NHL. It also discusses some of the plotting that can be done using this information. This is what an NHL player looks like courtesy of the NHL:
The above is some close-to-the-goal action between the goalie and offensive player of opposing teams. Similarly to soccer, the goalie in hockey can touch the puck with more of their body than can the average player, at least when they are right near the goal (which they usually but not always are)

The functions

NHL_records <- function(team_or_id = NULL, specifier = NULL){
## start with the URL that begins every API call
base_url <- "https://records.nhl.com/site/api"
## if the user inputs "franchise," append the base URL with the /franchise modifier to return the full historical dataset of team names
  if(identical(team_or_id, "franchise")){
    base_url <- paste0(base_url, "/franchise")
  }
  if(identical(team_or_id, "stats")){
    base_url <- paste0(base_url, "/franchise-team-totals")
  }
  ## mapping all of the team names to their respective franchise ID numbers
  if(identical(team_or_id, "New Jersey Devils")){
    team_or_id <- 23
  }
  if(identical(team_or_id, "New York Islanders")){
    team_or_id <- 22
  }
  if(identical(team_or_id, "New York Rangers")){
    team_or_id <- 10
  }
  if(identical(team_or_id, "Philadelphia Flyers")){
    team_or_id <- 16
  }
  if(identical(team_or_id, "Pittsburgh Penguins")){
    team_or_id <- 17
  }
  if(identical(team_or_id, "Boston Bruins")){
    team_or_id <- 6
  }
  if(identical(team_or_id, "Buffalo Sabres")){
    team_or_id <- 19
  }
  if(identical(team_or_id, "Montreal Canadiens")){
    team_or_id <- 1
  }
  if(identical(team_or_id, "Ottawa Senators")){
    team_or_id <- 30
  }
  if(identical(team_or_id, "Toronto Maple Leafs")){
    team_or_id <- 5
  }
  if(identical(team_or_id, "Carolina Hurricanes")){
    team_or_id <- 26
  }
  if(identical(team_or_id, "Florida Panthers")){
    team_or_id <- 33
  }
  if(identical(team_or_id, "Tampa Bay Lightning")){
    team_or_id <- 31
  }
  if(identical(team_or_id, "Washington Capitals")){
    team_or_id <- 24
  }
  if(identical(team_or_id, "Chicago Blackhawks")){
    team_or_id <- 11
  }
  if(identical(team_or_id, "Detroit Redwings")){
    team_or_id <- 12
  }
  if(identical(team_or_id, "Nashville Predators")){
    team_or_id <- 34
  }
  if(identical(team_or_id, "St. Lious Blues")){
    team_or_id <- 18
  }
  if(identical(team_or_id, "Calgary Flames")){
    team_or_id <- 21
  }
  if(identical(team_or_id, "Colorado Avalanche")){
    team_or_id <- 27
  }
  if(identical(team_or_id, "Edmonton Oilers")){
    team_or_id <- 25
  }
  if(identical(team_or_id, "Vancouver Canucks")){
    team_or_id <- 20
  }
  if(identical(team_or_id, "Anaheim Ducks")){
    team_or_id <- 32
  }
  if(identical(team_or_id, "Dallas Stars")){
    team_or_id <- 15
  }
  if(identical(team_or_id, "Los Angeles Kings")){
    team_or_id <- 14
  }
  if(identical(team_or_id, "San Jose Sharks")){
    team_or_id <- 29
  }
  if(identical(team_or_id, "Columbus Blue Jackets")){
    team_or_id <- 36
  }
  if(identical(team_or_id, "Minnesota Wild")){
    team_or_id <- 37
  }
  if(identical(team_or_id, "Winnipeg Jets")){
    team_or_id <- 35
  }
  if(identical(team_or_id, "Arizona Coyotes")){
    team_or_id <- 28
  }
  if(identical(team_or_id, "Vegas Golden Knights")){
    team_or_id <- 38
  }
  if(identical(team_or_id, "Seattle Kraken")){
    team_or_id <- 39
  }

##if the user only enters a number or team name
  if(is.numeric(team_or_id) && is.na(specifier)){
    base_url <- paste0(base_url, "/franchise-season-records?cayenneExp=franchiseId=", team_or_id)
  } 
## the user specifies goalie and skater records
  if(is.numeric(team_or_id) && identical(specifier, "goalie")){
    base_url <- paste0(base_url, "/franchise-goalie-records?cayenneExp=franchiseId=", team_or_id)
  }
  if(is.numeric(team_or_id) && identical(specifier, "skater")){
    base_url <- paste0(base_url, "/franchise-skater-records?cayenneExp=franchiseId=", team_or_id)
  }
#the user says admin to see retired numbers, etc.
  if(is.numeric(team_or_id) && identical(specifier, "admin")){
    base_url <- paste0(base_url, "/franchise-detail?cayenneExp=mostRecentTeamId=", team_or_id)
  }

  
## if the user puts in a word other than those specified above, return a message telling the user it was a bad input
## 
  #Make the API call and get the data into a nice form
  data_1 <- GET(base_url)
  data_2 <- content(data_1, "text", encoding = "UTF-8")
  data_3 <- fromJSON(data_2, flatten = TRUE)
  data_4 <- as.data.frame(data_3)
  data_5 <- tibble(data_4)
  return(data_5)
}

Similarly for this NHL stats API, one can simply put in the team ID to get a break down of a particular teams season stats. These stats are for the current season.

NHL_stats <- function(id){
  ##start with the base URL's that are the beginning and the end of all the API calls
base_url_2 <- "https://statsapi.web.nhl.com/api/v1/teams/"
ending_url <- "?expand=team.stats"
# put them together with the inputted user number in the middle
base_url_2 <- paste0(base_url_2, id, ending_url)
#need to allow user to request information for all the teams
if(identical(id, "all")){
  base_url_2 <- "https://statsapi.web.nhl.com/api/v1/teams?expand=team.stats"
}
#sequence of actions to take data from the API and return it in a nicer form
  data_1 <- GET(base_url_2)
  data_2 <- content(data_1, "text", encoding = "UTF-8")
  data_3 <- fromJSON(data_2, flatten = TRUE)
  data_4 <- as.data.frame(data_3)
  data_5 <- tibble(data_4)
  data_6 <- unnest(unnest(data_5,cols = c(teams.teamStats)),cols = c(splits))
  return(data_6)
}

In some cases, the output from NHL_stats or NHL_records is not in exactly the form you would expect, so I also built a function that simply pings the API and gets a tibble of data, without trying to unpack it in a particular way.

#the user simply indicates which API they are wanting to query with a modifier word
wrapper <- function(id, modifier){
  if(identical(modifier, "records")){
    return(NHL_records(id))
    }
  if(identical(modifier, "stats")){
    return(NHL_stats(id))
  }
  
  }

Pulling Data

Now that we have our underlying functions, NHL_records and NHL_stats, and our wrapper function that simply calls each of them given the appropriate input, it is time to actually start pulling some data. First I will grab the return for the stats API for the New Jersey Devils.

###start actually pulling some data

New_Jersey <- wrapper(1, "stats")
New_Jersey

## # A tibble: 2 x 65
##   copyright            teams.id teams.name   teams.link teams.abbreviat~ teams.teamName
##   <chr>                   <int> <chr>        <chr>      <chr>            <chr>         
## 1 NHL and the NHL Shi~        1 New Jersey ~ /api/v1/t~ NJD              Devils        
## 2 NHL and the NHL Shi~        1 New Jersey ~ /api/v1/t~ NJD              Devils        
## # ... with 59 more variables: teams.locationName <chr>, teams.firstYearOfPlay <chr>,
## #   stat.gamesPlayed <int>, stat.wins <chr>, stat.losses <chr>, stat.ot <chr>,
## #   stat.pts <chr>, stat.ptPctg <chr>, stat.goalsPerGame <chr>,
## #   stat.goalsAgainstPerGame <chr>, stat.evGGARatio <chr>,
## #   stat.powerPlayPercentage <chr>, stat.powerPlayGoals <chr>,
## #   stat.powerPlayGoalsAgainst <chr>, stat.powerPlayOpportunities <chr>,
## #   stat.penaltyKillPercentage <chr>, stat.shotsPerGame <chr>,
## #   stat.shotsAllowed <chr>, stat.winScoreFirst <chr>, stat.winOppScoreFirst <chr>,
## #   stat.winLeadFirstPer <chr>, stat.winLeadSecondPer <chr>,
## #   stat.winOutshootOpp <chr>, stat.winOutshotByOpp <chr>, stat.faceOffsTaken <chr>,
## #   stat.faceOffsWon <chr>, stat.faceOffsLost <chr>, stat.faceOffWinPercentage <chr>,
## #   stat.shootingPctg <dbl>, stat.savePctg <dbl>, stat.penaltyKillOpportunities <chr>,
## #   stat.savePctRank <chr>, stat.shootingPctRank <chr>, team.id <int>,
## #   team.name <chr>, team.link <chr>, type.displayName <chr>, type.gameType.id <chr>,
## #   type.gameType.description <chr>, type.gameType.postseason <lgl>,
## #   teams.shortName <chr>, teams.officialSiteUrl <chr>, teams.franchiseId <int>,
## #   teams.active <lgl>, teams.venue.name <chr>, teams.venue.link <chr>,
## #   teams.venue.city <chr>, teams.venue.timeZone.id <chr>,
## #   teams.venue.timeZone.offset <int>, teams.venue.timeZone.tz <chr>,
## #   teams.division.id <int>, teams.division.name <chr>, teams.division.link <chr>,
## #   teams.conference.id <int>, teams.conference.name <chr>,
## #   teams.conference.link <chr>, teams.franchise.franchiseId <int>,
## #   teams.franchise.teamName <chr>, teams.franchise.link <chr>

This output has all kinds of statistics for this season for the New Jersey Devils including losses, goals and even derived statistics such as goals per game. The problem with this is that we only have information for New Jersey!

What I would like to do with these functions is compare stats across teams and find some correlations and interesting representations of data. To do this we will call our wrapper function, specify “stats,” and request “all” to get data for all the teams. The function does a great job unpacking the data (it begins with a doubly nested data object), but it does leave in the rankings for each stat for the 31 teams. I just want the raw numbers, so I will filter out all rows for which “games played” is na. This is an effective way to do it because for all of the rows that contain rank information, the stat for games played is na, whereas for the raw information the games played stat has what you would expect: the number of games each team has played.

motherload <- wrapper("all", "stats")
head(motherload)

## # A tibble: 6 x 66
##   copyright            teams.id teams.name   teams.link teams.abbreviat~ teams.teamName
##   <chr>                   <int> <chr>        <chr>      <chr>            <chr>         
## 1 NHL and the NHL Shi~        1 New Jersey ~ /api/v1/t~ NJD              Devils        
## 2 NHL and the NHL Shi~        1 New Jersey ~ /api/v1/t~ NJD              Devils        
## 3 NHL and the NHL Shi~        2 New York Is~ /api/v1/t~ NYI              Islanders     
## 4 NHL and the NHL Shi~        2 New York Is~ /api/v1/t~ NYI              Islanders     
## 5 NHL and the NHL Shi~        3 New York Ra~ /api/v1/t~ NYR              Rangers       
## 6 NHL and the NHL Shi~        3 New York Ra~ /api/v1/t~ NYR              Rangers       
## # ... with 60 more variables: teams.locationName <chr>, teams.firstYearOfPlay <chr>,
## #   stat.gamesPlayed <int>, stat.wins <chr>, stat.losses <chr>, stat.ot <chr>,
## #   stat.pts <chr>, stat.ptPctg <chr>, stat.goalsPerGame <chr>,
## #   stat.goalsAgainstPerGame <chr>, stat.evGGARatio <chr>,
## #   stat.powerPlayPercentage <chr>, stat.powerPlayGoals <chr>,
## #   stat.powerPlayGoalsAgainst <chr>, stat.powerPlayOpportunities <chr>,
## #   stat.penaltyKillPercentage <chr>, stat.shotsPerGame <chr>,
## #   stat.shotsAllowed <chr>, stat.winScoreFirst <chr>, stat.winOppScoreFirst <chr>,
## #   stat.winLeadFirstPer <chr>, stat.winLeadSecondPer <chr>,
## #   stat.winOutshootOpp <chr>, stat.winOutshotByOpp <chr>, stat.faceOffsTaken <chr>,
## #   stat.faceOffsWon <chr>, stat.faceOffsLost <chr>, stat.faceOffWinPercentage <chr>,
## #   stat.shootingPctg <dbl>, stat.savePctg <dbl>, stat.penaltyKillOpportunities <chr>,
## #   stat.savePctRank <chr>, stat.shootingPctRank <chr>, team.id <int>,
## #   team.name <chr>, team.link <chr>, type.displayName <chr>, type.gameType.id <chr>,
## #   type.gameType.description <chr>, type.gameType.postseason <lgl>,
## #   teams.shortName <chr>, teams.officialSiteUrl <chr>, teams.franchiseId <int>,
## #   teams.active <lgl>, teams.venue.name <chr>, teams.venue.link <chr>,
## #   teams.venue.city <chr>, teams.venue.id <int>, teams.venue.timeZone.id <chr>,
## #   teams.venue.timeZone.offset <int>, teams.venue.timeZone.tz <chr>,
## #   teams.division.id <int>, teams.division.name <chr>, teams.division.link <chr>,
## #   teams.conference.id <int>, teams.conference.name <chr>,
## #   teams.conference.link <chr>, teams.franchise.franchiseId <int>,
## #   teams.franchise.teamName <chr>, teams.franchise.link <chr>

#taking out the rows that contain rank information
motherload <- motherload %>% filter(!is.na(stat.gamesPlayed))
head(motherload)

## # A tibble: 6 x 66
##   copyright            teams.id teams.name   teams.link teams.abbreviat~ teams.teamName
##   <chr>                   <int> <chr>        <chr>      <chr>            <chr>         
## 1 NHL and the NHL Shi~        1 New Jersey ~ /api/v1/t~ NJD              Devils        
## 2 NHL and the NHL Shi~        2 New York Is~ /api/v1/t~ NYI              Islanders     
## 3 NHL and the NHL Shi~        3 New York Ra~ /api/v1/t~ NYR              Rangers       
## 4 NHL and the NHL Shi~        4 Philadelphi~ /api/v1/t~ PHI              Flyers        
## 5 NHL and the NHL Shi~        5 Pittsburgh ~ /api/v1/t~ PIT              Penguins      
## 6 NHL and the NHL Shi~        6 Boston Brui~ /api/v1/t~ BOS              Bruins        
## # ... with 60 more variables: teams.locationName <chr>, teams.firstYearOfPlay <chr>,
## #   stat.gamesPlayed <int>, stat.wins <chr>, stat.losses <chr>, stat.ot <chr>,
## #   stat.pts <chr>, stat.ptPctg <chr>, stat.goalsPerGame <chr>,
## #   stat.goalsAgainstPerGame <chr>, stat.evGGARatio <chr>,
## #   stat.powerPlayPercentage <chr>, stat.powerPlayGoals <chr>,
## #   stat.powerPlayGoalsAgainst <chr>, stat.powerPlayOpportunities <chr>,
## #   stat.penaltyKillPercentage <chr>, stat.shotsPerGame <chr>,
## #   stat.shotsAllowed <chr>, stat.winScoreFirst <chr>, stat.winOppScoreFirst <chr>,
## #   stat.winLeadFirstPer <chr>, stat.winLeadSecondPer <chr>,
## #   stat.winOutshootOpp <chr>, stat.winOutshotByOpp <chr>, stat.faceOffsTaken <chr>,
## #   stat.faceOffsWon <chr>, stat.faceOffsLost <chr>, stat.faceOffWinPercentage <chr>,
## #   stat.shootingPctg <dbl>, stat.savePctg <dbl>, stat.penaltyKillOpportunities <chr>,
## #   stat.savePctRank <chr>, stat.shootingPctRank <chr>, team.id <int>,
## #   team.name <chr>, team.link <chr>, type.displayName <chr>, type.gameType.id <chr>,
## #   type.gameType.description <chr>, type.gameType.postseason <lgl>,
## #   teams.shortName <chr>, teams.officialSiteUrl <chr>, teams.franchiseId <int>,
## #   teams.active <lgl>, teams.venue.name <chr>, teams.venue.link <chr>,
## #   teams.venue.city <chr>, teams.venue.id <int>, teams.venue.timeZone.id <chr>,
## #   teams.venue.timeZone.offset <int>, teams.venue.timeZone.tz <chr>,
## #   teams.division.id <int>, teams.division.name <chr>, teams.division.link <chr>,
## #   teams.conference.id <int>, teams.conference.name <chr>,
## #   teams.conference.link <chr>, teams.franchise.franchiseId <int>,
## #   teams.franchise.teamName <chr>, teams.franchise.link <chr>

I also would like to look at some historical statistics, which we again will pull using our wrapper function.

#"stats" is used as a modifier and as input into records
franchises <- wrapper("stats", "records")

The Graphs

The first thing I would like to look at is how well scoring correlates with winning. To look at that I am going to bring up a scatterplot of wins and goals.

scatter <- ggplot(motherload, aes(stat.wins, stat.goalsPerGame,))
scatter + geom_jitter(color = "green") + xlab("Wins") + ylab("Goals Per Game") + ggtitle("The Correlation Between Wins and Goals Per Game")

As expected, we can see that wins and goals per game are highly correlated. In fact, they are even more highly correlated than I would have thought a priori. In basketball and football, for instance, it is frequently the case that there is a great offensive team with a poor overall record. This is not the case here as is easily seen by the fact that there are no dots high up on the graph that appear on the left half. We can see that, in fact, the top four winningest teams are in the top five in scoring! It would appear in hockey that the best offense is a good offense, and perhaps even like the old adage the best defense is a good offense!

scatter2 <- ggplot(motherload, aes(stat.wins, stat.goalsAgainstPerGame))
scatter2 + geom_point(color = "darkred") + xlab("Wins") + ylab("Goals Allowed") + ggtitle("Allowing Goals Leads to Loses")

Of course we don’t want to let offense have all the fun! We look at the other side of the puck with this scatter plot as we plot wins versus the number of goals a team’s opponent scores. I expected these two values to be negatively correlated, which they are. I would make the case, though, that in fact offense predominates in hockey because there are a few very good defensive teams that are not elite from a total wins perspective.

Now that I have a little bit of an understanding of how offense and defense impact winning in hockey, I would like to get some historical understanding of the different franchises and thus the history of the league. I would like to look at a boxplot of the total games played by each franchise.

box <- ggplot(franchises, aes(data.gamesPlayed))
box + geom_boxplot(color = "orange", fill = "black") + xlab("Games Played All-Time") + ggtitle("Six Franchises Have Played More Games than the Rest")

This is very interesting to look at! One can instantly understand a little bit about the history of the NHL from one simple graph. Most franchises in the NHL have had less than 2000 games in their history, while there are six that have over six thousand! This makes me wonder if those six franchises are the founding members of the NHL. There are two statistics missing from the season statistics that I am interested in: face off’s taken per game, and power play opportunities per game. I will create these new variables by appending them to my table called “motherload.”

## create new per game stats by dividing by games played. Note that power play opportunities and face offs taken, though they are numbers, are coded as characters so I will need to coerce it to an int.
motherload$stat.powerPlayOpportunities <- as.numeric(motherload$stat.powerPlayOpportunities)

motherload$stat.gamesPlayed <- as.numeric(motherload$stat.gamesPlayed)

motherload$stat.faceOffsTaken <- as.numeric(motherload$stat.faceOffsTaken)
#now actually calculate our new statistics
face_power <- motherload %>% mutate(faceoffs_perGame = stat.faceOffsTaken / stat.gamesPlayed)

face_power <- face_power %>% mutate(powerplays_pergame = stat.powerPlayOpportunities/ stat.gamesPlayed)

head(face_power)

## # A tibble: 6 x 68
##   copyright            teams.id teams.name   teams.link teams.abbreviat~ teams.teamName
##   <chr>                   <int> <chr>        <chr>      <chr>            <chr>         
## 1 NHL and the NHL Shi~        1 New Jersey ~ /api/v1/t~ NJD              Devils        
## 2 NHL and the NHL Shi~        2 New York Is~ /api/v1/t~ NYI              Islanders     
## 3 NHL and the NHL Shi~        3 New York Ra~ /api/v1/t~ NYR              Rangers       
## 4 NHL and the NHL Shi~        4 Philadelphi~ /api/v1/t~ PHI              Flyers        
## 5 NHL and the NHL Shi~        5 Pittsburgh ~ /api/v1/t~ PIT              Penguins      
## 6 NHL and the NHL Shi~        6 Boston Brui~ /api/v1/t~ BOS              Bruins        
## # ... with 62 more variables: teams.locationName <chr>, teams.firstYearOfPlay <chr>,
## #   stat.gamesPlayed <dbl>, stat.wins <chr>, stat.losses <chr>, stat.ot <chr>,
## #   stat.pts <chr>, stat.ptPctg <chr>, stat.goalsPerGame <chr>,
## #   stat.goalsAgainstPerGame <chr>, stat.evGGARatio <chr>,
## #   stat.powerPlayPercentage <chr>, stat.powerPlayGoals <chr>,
## #   stat.powerPlayGoalsAgainst <chr>, stat.powerPlayOpportunities <dbl>,
## #   stat.penaltyKillPercentage <chr>, stat.shotsPerGame <chr>,
## #   stat.shotsAllowed <chr>, stat.winScoreFirst <chr>, stat.winOppScoreFirst <chr>,
## #   stat.winLeadFirstPer <chr>, stat.winLeadSecondPer <chr>,
## #   stat.winOutshootOpp <chr>, stat.winOutshotByOpp <chr>, stat.faceOffsTaken <dbl>,
## #   stat.faceOffsWon <chr>, stat.faceOffsLost <chr>, stat.faceOffWinPercentage <chr>,
## #   stat.shootingPctg <dbl>, stat.savePctg <dbl>, stat.penaltyKillOpportunities <chr>,
## #   stat.savePctRank <chr>, stat.shootingPctRank <chr>, team.id <int>,
## #   team.name <chr>, team.link <chr>, type.displayName <chr>, type.gameType.id <chr>,
## #   type.gameType.description <chr>, type.gameType.postseason <lgl>,
## #   teams.shortName <chr>, teams.officialSiteUrl <chr>, teams.franchiseId <int>,
## #   teams.active <lgl>, teams.venue.name <chr>, teams.venue.link <chr>,
## #   teams.venue.city <chr>, teams.venue.id <int>, teams.venue.timeZone.id <chr>,
## #   teams.venue.timeZone.offset <int>, teams.venue.timeZone.tz <chr>,
## #   teams.division.id <int>, teams.division.name <chr>, teams.division.link <chr>,
## #   teams.conference.id <int>, teams.conference.name <chr>,
## #   teams.conference.link <chr>, teams.franchise.franchiseId <int>,
## #   teams.franchise.teamName <chr>, teams.franchise.link <chr>,
## #   faceoffs_perGame <dbl>, powerplays_pergame <dbl>

I would like to now take our newly created powerplays_pergame statistic and plot it on a bar plot to see the different teams side by side.

##create plot and add bar plot layer
bar <- ggplot(face_power, aes(powerplays_pergame))
bar + geom_bar(color = "blue") + xlab("Powerplays per Game") + ylab("Number") + ggtitle("Team Having Most Powerplays Has Nearly Twice as Many as the Team with Least")

Wow check this out! Most teams have between 2.5 and 3.25 Powerplays per game but the one with the most has going on 4 Powerplays per game. The fact that on the professional level, one team can generate nearly twice the Powerplays of another is interesting. Find out what a Powerplay is below.

I don’t know a whole lot about hockey, but one thing I have picked up from the few games I have watched is that Powerplay goals are very important. This is when one team has one more player on the ice than the other because of a foul one team committed. Let’s look at a histogram showing the number of teams with different ranges of Powerplay goals.

motherload$stat.powerPlayGoals <- as.numeric(motherload$stat.powerPlayGoals)
hist <- ggplot(motherload, aes(stat.powerPlayGoals))
hist + geom_histogram(binwidth = 4, color = "purple", fill = "yellow") + xlab("Number Powerplay Goals") + ylab("Number of Teams With") + ggtitle("Most Teams Have Roughly 30 to 40 Powerplay Goals this Season")

You can see that some teams have as few as 10 power play goals while others have about 50 and most teams have around 30. Once again, a range of 10-50 for any statistic in sports strikes me as being highly variable - one team has 5 times the power play goals!

I’d like to explore how many franchises are active NHL franchises and how many franchises are historical/previously active NHL franchises.

franchises$data.activeFranchise <- as.factor(franchises$data.activeFranchise)
table(franchises$data.activeFranchise)

## 
##  0  1 
## 18 87

Whether the database intends to be using 0 as active or 1 as active, something has gone awry here because I know there are 31 active franchises, hence more than 18 and less than 87. Looking at the data it appears that 1 is meant to be active and 0 is meant to be historical, the problem is just that we have some confusing duplication in the data.

One statistic that is bound to matter a lot in hockey is shooting percentage, what percentage of shots go in. As a way of looking how that varies between teams I am going to look at a numerical summary of it.

summ <- summary(motherload$stat.shootingPctg)
summ

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   8.200   9.000   9.500   9.665  10.350  11.500

We can see that the lowest shooting percentage is 8.2 and the highest is 11.5 with a mean of 9.665. To think that 33 more shots out of 1000 go in for the most accurate team than the least accurate, that is a big difference in number of goals made per thousand shots!