In RobWHickman/CamStroke:

library(tidyverse)
library(rvest)
library(magrittr)
library(lubridate)

fairbairns_2019 <- "https://jcbc.jesus.cam.ac.uk/old/fairbairns/results.php" %>%
  read_html()

men_2019 <- fairbairns_2019 %>%
  html_nodes(xpath = '//*[@id="content"]/table[6]') %>%
      html_table() %>%
      as.data.frame() %>%
      `colnames<-`(c("position", "race_no", "crew", "time", "winner_diff", "half_time", "winner_half_diff")) %>%
      dplyr::filter(position != "Position") %>%
      mutate(year = 2019) %>%
      mutate(gender = "M") %>%
      select(position, crew, gender, time, year)
women_2019 <- fairbairns_2019 %>%
  html_nodes(xpath = '//*[@id="content"]/table[7]') %>%
      html_table() %>%
      as.data.frame() %>%
      `colnames<-`(c("position", "race_no", "crew", "time", "winner_diff", "half_time", "winner_half_diff")) %>%
      dplyr::filter(position != "Position") %>%
      mutate(year = 2019) %>%
      mutate(gender = "F") %>%
      select(position, crew, gender, time, year)

fairbairns_2019_data <- rbind(men_2019, women_2019)


fairbairns_data <- paste0("https://jcbc.jesus.cam.ac.uk/old/fairbairns/archive/results.php?year=", 2010:2017) %>%
  lapply(., function(year) {
    men <- read_html(year) %>%
      html_nodes(xpath = '//*[@id="content"]/table[3]') %>%
      html_table() %>%
      as.data.frame() %>%
      `colnames<-`(c("position", "race_no", "crew", "time", "winner_diff", "half_time", "winner_half_diff")) %>%
      dplyr::filter(position != "Position") %>%
      mutate(year = as.numeric(gsub(".*=", "", year))+1) %>%
      mutate(gender = "M") %>%
      select(position, crew, gender, time, year)
    women <- read_html(year) %>%
      html_nodes(xpath = '//*[@id="content"]/table[4]') %>%
      html_table() %>%
      as.data.frame() %>%
      `colnames<-`(c("position", "race_no", "crew", "time", "winner_diff", "half_time", "winner_half_diff")) %>%
      dplyr::filter(position != "Position") %>%
      mutate(year = as.numeric(gsub(".*=", "", year))+1) %>%
      mutate(gender = "F") %>%
      select(position, crew, gender, time, year)
    all <- rbind(men, women)
  }) %>%
  do.call(rbind, .) %>%
  rbind(., fairbairns_2019_data) %>%
  mutate(Fairbairns_Name = gsub(" I.*| V.*", "", crew))  %>%
  mutate(college = case_when(
    grepl("Caius", Fairbairns_Name) ~ "Caius",
    grepl("St Cat", Fairbairns_Name) ~ "Catz",
    grepl("Christ's", Fairbairns_Name) ~ "Christs",
    grepl("Churchill", Fairbairns_Name) ~ "Churchill",
    grepl("Clare Hall", Fairbairns_Name) ~ "ClareHall",
    grepl("Clare", Fairbairns_Name) ~ "Clare",
    grepl("Corpus Christi", Fairbairns_Name) ~ "Corpus",
    grepl("Darwin", Fairbairns_Name) ~ "Darwin",
    grepl("Downing", Fairbairns_Name) ~ "Downing",
    grepl("Emmanuel", Fairbairns_Name) ~ "Emma",
    grepl("First and Third Trinity", Fairbairns_Name) ~ "FaT",
    grepl("Fitzwilliam", Fairbairns_Name) ~ "Fitz",
    grepl("Girton", Fairbairns_Name) ~ "Girton",
    grepl("Homerton", Fairbairns_Name) ~ "Homerton",
    grepl("Hughes Hall$", Fairbairns_Name) ~ "HughesHall",
    Fairbairns_Name == "Hughes Hall/ Lucy Cavendish" ~ "HughesHall/LucyCav",
    Fairbairns_Name == "Lucy Cavendish" ~ "LucyCav",
    Fairbairns_Name == "Newnham" ~ "Newnham",
    Fairbairns_Name == "Jesus" ~ "Jesus",
    grepl("King", Fairbairns_Name) ~ "Kings",
    grepl("Margaret$", Fairbairns_Name) ~ "Maggie",
    grepl("^Magdalene", Fairbairns_Name) ~ "Magdalene",
    grepl("Pembroke$", Fairbairns_Name) ~ "Pembroke",
    Fairbairns_Name == "Queen's" ~ "Queens",
    Fairbairns_Name == "Queens" ~ "Queens",
    grepl("Peterhouse", Fairbairns_Name) ~ "Peterhouse",
    grepl("Robinson", Fairbairns_Name) ~ "Robinson",
    Fairbairns_Name == "Selwyn" ~ "Selwyn",
    grepl("Sidney Sussex", Fairbairns_Name) ~ "Sidney",
    grepl("St Edmund's", Fairbairns_Name) ~ "Eddies",
    grepl("Trinity Hall", Fairbairns_Name) ~ "TitHall",
    grepl("Wolfson", Fairbairns_Name) ~ "Wolfson",
    grepl("Wolfson/Darwin", Fairbairns_Name) ~ "Wolfson/Darwin"
  )) %>%
  filter(!is.na(college)) %>%
  mutate(seconds = as.numeric(ms(time))) %>%
  mutate(crew = utils:::.roman2numeric(trimws(gsub("(.*)( I.*| V.*)", "\\2", crew)))) %>%
  mutate(crew = ifelse(is.na(crew), 1, crew)) %>%
  select(year, college, crew, gender, seconds) %>%
  mutate(race = "Fairbairns")

saveRDS(fairbairns_data, "fairbairns_results.rds")

bumps_data <- read.csv("C:/Users/WS-Guest/Documents/rob_r/RoweR/bumps.csv",
                       stringsAsFactors = FALSE) %>%
  dplyr::filter(Year > 2013, Gender == "M") %>%
  setDT() %>%
  melt.data.table(id.vars = c("College", "Crew", "Year", "Gender", "StartPos"),
                 value.name = "change",
                 variable.name = "day") %>%
  .[is.na(change), change := 0] %>%
  .[, cum_change := cumsum(change), by = c("College", "Crew", "Year")] %>%
  .[, day_start_pos := StartPos - cum_change + change] %>%
  .[day_start_pos < 30]

college_names <- read.csv("C:/Users/WS-Guest/Documents/rob_r/RoweR/college_names.csv",
                       stringsAsFactors = FALSE)

fairbairns_data %<>%
  left_join(.,
            select(college_names, College, Fairbairns_Name),
            by = "Fairbairns_Name") %>%
  dplyr::filter(!is.na(College)) %>%
  mutate(Crew = utils:::.roman2numeric(trimws(gsub("(.*)( I.*| V.*)", "\\2", crew)))) %>%
  select(College, Crew, Year = year, f_time = time)

bumps_data %<>% left_join(., fairbairns_data, by = c("College", "Crew", "Year"))

bumps_data$f_time <- as.numeric(lubridate::ms(bumps_data$f_time))

bumps_data2 <- bumps_data %>%
  setDT() %>%
  .[order(Year, day, StartPos)] %>%
  .[, f_time_diff := f_time - lead(f_time), by = c("Year", "day")] %>%
  .[!is.na(f_time_diff)] %>%
  .[change < 0, bumped := 1] %>%
  .[change >= 0, bumped := 0]

binomial_smooth <- function(...) {
  geom_smooth(method = "glm", method.args = list(family = "binomial"), ...)
}

p <- ggplot(data = bumps_data2, aes(y = bumped, x = f_time_diff)) +
  geom_point(alpha = 0.4) +
  xlab("second slower at Fairbairns than Chasing Boat") +
  ggtitle("Whether Div2 and Above Boats get bumped or not based on Previous Fairbairns Results of Them/Chasing Boat") +
  binomial_smooth()

RobWHickman/CamStroke documentation built on Nov. 22, 2019, 12:04 a.m.