Reproducible Scores Code

gistfile1.txt

# Activate these packages
library(shiny)
library(DT)
library(shinythemes)
library(rvest)
library(expss)
library(dplyr)
library(tidyr)
library(stringr)
library(sqldf)
library(scales)

# List of schools and class: 
## https://nsaa-static.s3.amazonaws.com/textfile/bask/bbbclass.pdf

# Manually create Class A names and enrollment dataframe
School <-
  c(
    'Omaha South',
    'Omaha Central',
    'Grand Island',
    'Millard North',
    'Millard South',
    'Lincoln High',
    'Lincoln East',
    'Millard West',
    'North Star',
    'Omaha North',
    'Creighton Prep',
    'Lincoln Southeast',
    'Lincoln Southwest',
    'Burke',
    'Omaha Westside',
    'Papillion-LaVista South',
    'Bryan',
    'Papillion-LaVista',
    'Omaha Northwest',
    'Lincoln Northeast',
    'Kearney',
    'Bellevue West',
    'Bellevue East',
    'Gretna',
    'Fremont',
    'Elkhorn South',
    'Benson',
    'Norfolk',
    'Columbus',
    'Pius X',
    'North Platte'
  )

Enrollment <-
  c(
    2226,
    2094,
    1970,
    1923,
    1903,
    1754,
    1734,
    1687,
    1623,
    1584,
    1562,
    1557,
    1557,
    1547,
    1462,
    1447,
    1439,
    1381,
    1316,
    1278,
    1215,
    1145,
    1123,
    1121,
    1113,
    1076,
    1076,
    1038,
    965,
    898,
    896
  )

classA <- data.frame(School, Enrollment)

# create game scores table -- I added game dates manually
maxprep_baseURL <- 
  "https://www.maxpreps.com/list/schedules_scores.aspx?date="

maxprep_paramURL <- 
  "&ssid=8f5a3f82-d7d8-4f76-a137-3c884eae5b54&statedivisionid=07ba312e-1587-4455-b373-ca343d52f4b2"

game_dates <- c(
  "12/3/2020",
  "12/4/2020",
  "12/5/2020",
  "12/7/2020",
  "12/10/2020",
  "12/11/2020",
  "12/12/2020",
  "12/15/2020",
  "12/16/2020",
  "12/17/2020",
  "12/18/2020",
  "12/19/2020",
  "12/21/2020",
  "12/22/2020",
  "12/28/2020",
  "12/29/2020",
  "12/31/2020",
  "1/2/2021",
  "1/5/2021",
  "1/7/2021",
  "1/8/2021",
  "1/9/2021",
  "1/12/2021",
  "1/14/2021",
  "1/15/2021",
  "1/16/2021",
  "1/19/2021",
  "1/21/2021",
  "1/22/2021",
  "1/23/2021",
  "1/26/2021",
  "1/28/2021",
  "1/29/2021",
  "1/30/2021",
  "2/2/2021",
  "2/4/2021",
  "2/5/2021",
  "2/6/2021",
  "2/9/2021",
  "2/11/2021",
  "2/12/2021",
  "2/13/2021",
  "2/16/2021",
  "2/18/2021",
  "2/19/2021",
  "2/20/2021",
  "2/26/2021",
  "2/27/2021",
  "3/1/2021",
  "3/2/2021",
  "3/11/2021",
  "3/12/2021",
  "3/13/2021"
)

maxprep_page_list <- 
  as.list(paste0(maxprep_baseURL, game_dates, maxprep_paramURL))

maxprep_html <- lapply(maxprep_page_list, FUN=function(URLLink){
  read_html(URLLink) %>% html_nodes("[data-contest-state='boxscore']") %>% html_text()
})

# Unlist
scores <-
  unlist(maxprep_html)

scores <- 
  gsub("Final","", scores)

scores <- 
  grep("#", scores, invert = TRUE, value = TRUE)

scores <- 
  data.frame(scores)

colnames(scores) <- 
  c("V1")

scores <- 
  scores %>%
  mutate(V1 = gsub("(\\d+)", ";\\1;", V1)) %>%
  separate(V1, c(NA, "No1", "Let1", "No2", "Let2"), sep = " *; *")

colnames(scores) <- 
  c("Away_Score", "Away_Team", "Home_Score", "Home_Team")

scores$Winner <- 
  ifelse(scores$Away_Score > scores$Home_Score, scores$Away_Team, scores$Home_Team)