R/split_stats_functions.R
In ryurko/nflWAR: Computes WAR for Offensive Players in the NFL
#' Join Various Statistics to the Position Tables
#'
#' @param model_data_list List that must have the following:
#' \itemize{
#' \item{"pass_model_df"} - Passing play-by-play data
#' \item{"rush_model_df"} - Rushing play-by-play data
#' \item{"QB_table"} - Table of QBs
#' \item{"RB_table"} - Table of RBs
#' \item{"WR_table"} - Table of WRs
#' \item{"TE_table"} - Table of TEs
#' }
#' @return The input position tables with all the different
#' types of statistics included.
#' @examples
#' # Calculate the various types of statistics for each position:
#' model_data_list <- model_data_list %>%
#' join_position_statistics(model_data_list)
#' @export
join_position_statistics <- function(model_data_list) {
# Use the split functions to calculate the different
# types of statistics:
passing_stats <- calc_passing_splits(model_data_list$pass_model_df)
receiving_stats <- calc_receiving_splits(model_data_list$pass_model_df)
rushing_stats <- calculate_rushing_splits(model_data_list$rush_model_df)
# Drop the redundant columns in each dataset:
passing_stats <- passing_stats %>%
dplyr::select(-Pass_Attempts)
receiving_stats <- receiving_stats %>%
dplyr::select(-Targets)
rushing_stats <- rushing_stats %>%
dplyr::select(-Rushes, -Sacks)
# Now join the stats to the position tables, and return:
model_data_list$QB_table <- model_data_list$QB_table %>%
dplyr::left_join(passing_stats,
by = c("Player_ID_Name" = "Passer_ID_Name")) %>%
dplyr::left_join(rushing_stats,
by = c("Player_ID_Name" = "Rusher_ID_Name")) %>%
dplyr::mutate(AdjNetYardsAtt = (Pass_Yards_Gained + 20 * Pass_TDs - 45 * INTs + Sack_Yards_Lost) / (Pass_Attempts + Sacks))
model_data_list$QB_table[is.na(model_data_list$QB_table)] <- 0
# Expression for the non-QB tables:
join_nonQB_stats <- . %>%
dplyr::left_join(receiving_stats,
by = c("Player_ID_Name" = "Receiver_ID_Name")) %>%
dplyr::left_join(rushing_stats,
by = c("Player_ID_Name" = "Rusher_ID_Name"))
model_data_list$WR_table <- model_data_list$WR_table %>% join_nonQB_stats
model_data_list$WR_table[is.na(model_data_list$WR_table)] <- 0
model_data_list$RB_table <- model_data_list$RB_table %>% join_nonQB_stats
model_data_list$RB_table[is.na(model_data_list$RB_table)] <- 0
model_data_list$TE_table <- model_data_list$TE_table %>% join_nonQB_stats
model_data_list$TE_table[is.na(model_data_list$TE_table)] <- 0
return(model_data_list)
}
#' Calculate Passing Statistics for Given Splits
#'
#' @param pass_pbp Data frame of passing plays.
#' @param splits String vector of variables to split on, default
#' is "Passer_ID_Name" (splits are levels to calculate the
#' statistics for).
#' @return Data frame of passing statistics evaluated at the
#' given level of splits.
#' @examples
#' # Calculate passing statistics :
#' model_data_list <- calculate_passing_splits(model_data_list$pass_model_df,
#' "Passer_ID_Name")
#' @export
calc_passing_splits <- function(pass_pbp, splits = "Passer_ID_Name") {
# Turn the splits into the group variables:
split_groups <- lapply(splits, as.symbol)
# Create the Game_Drive column:
pass_pbp %>%
dplyr::mutate(Game_Drive = paste(as.character(GameID),
as.character(Drive), sep = "-")) %>%
# Group by the given splits:
dplyr::group_by_(.dots = split_groups) %>%
# Calculate the passing statistics at the given splits level:
dplyr::summarise(Pass_Attempts = n(),
Completions = sum(Reception, na.rm = TRUE),
Pass_Drives = n_distinct(Game_Drive),
Comp_Perc = Completions / Pass_Attempts,
Pass_Yards_Gained = sum(Yards.Gained, na.rm = TRUE),
Total_AirYards = sum(AirYards, na.rm = TRUE),
Total_AirYards_Comp = sum(Reception*AirYards, na.rm = TRUE),
Total_YardsAfterCatch = sum(Reception*YardsAfterCatch, na.rm = TRUE),
Yards_per_Att = Pass_Yards_Gained / Pass_Attempts,
Yards_per_Comp = Pass_Yards_Gained / Completions,
Yards_per_Drive = Pass_Yards_Gained / Pass_Drives,
AirYards_per_Att = Total_AirYards / Pass_Attempts,
AirYards_per_Comp = Total_AirYards / Completions,
AirYards_per_Drive = Total_AirYards / Pass_Drives,
AirYards_Comp_per_Att = Total_AirYards_Comp / Pass_Attempts,
AirYards_Comp_per_Comp = Total_AirYards_Comp / Completions,
AirYards_Comp_per_Drive = Total_AirYards_Comp / Pass_Drives,
YardsAfterCatch_per_Att = Total_YardsAfterCatch / Pass_Attempts,
YardsAfterCatch_per_Comp = Total_YardsAfterCatch / Completions,
YardsAfterCatch_per_Drive = Total_YardsAfterCatch / Pass_Drives,
PACR = Pass_Yards_Gained / Total_AirYards,
Times_Hit = sum(QBHit, na.rm = TRUE),
Times_Hit_per_Att = Times_Hit / Pass_Attempts,
Times_Hit_per_Comp = Times_Hit / Completions,
Times_Hit_per_Drive = Times_Hit / Pass_Drives,
INTs = sum(InterceptionThrown, na.rm = TRUE),
Pass_TDs = sum(Touchdown, na.rm = TRUE),
Air_TDs = sum(as.numeric(YardsAfterCatch == 0) * Touchdown, na.rm = TRUE),
Air_TD_Rate = Air_TDs / Pass_TDs,
TD_to_INT = Pass_TDs / INTs,
TD_per_Att = Air_TDs / Pass_Attempts,
Air_TD_per_Att = Air_TDs / Pass_Attempts,
TD_per_Comp = Pass_TDs / Completions,
Air_TD_per_Comp = Air_TDs / Completions,
TD_per_Drive = Pass_TDs / Pass_Drives,
Air_TD_per_Drive = Air_TDs / Pass_Drives,
INTs_per_Att = INTs / Pass_Attempts,
Passer_Rating = ((((Comp_Perc - .3)*5) + ((Yards_per_Att - 3)*.25) + (TD_per_Att*20) + (2.375 - (INTs_per_Att * 25))) / 6) * 100,
INTs_per_Comp = INTs / Completions,
INTs_per_Drive = INTs / Pass_Drives,
Pass_EPA = sum(EPA, na.rm = TRUE),
Pass_EPA_per_Att = Pass_EPA / Pass_Attempts,
Pass_EPA_per_Comp = Pass_EPA / Completions,
Pass_EPA_per_Drive = Pass_EPA / Pass_Drives,
Pass_Success_Rate = length(which(EPA > 0)) / Pass_Attempts,
Comp_EPA = sum(Reception * EPA, na.rm = TRUE),
Comp_EPA_per_Att = Comp_EPA / Pass_Attempts,
Comp_EPA_per_Comp = Comp_EPA / Completions,
Comp_EPA_per_Drive = Comp_EPA / Pass_Drives,
EPA_Comp_Perc = Comp_EPA / sum(abs(EPA), na.rm = TRUE),
Pass_WPA = sum(WPA, na.rm = TRUE),
Pass_WPA_per_Att = Pass_WPA / Pass_Attempts,
Pass_WPA_per_Comp = Pass_WPA / Completions,
Pass_WPA_per_Drive = Pass_WPA / Pass_Drives,
Pass_Win_Rate = length(which(WPA > 0)) / Pass_Attempts,
Comp_WPA = sum(Reception * WPA, na.rm = TRUE),
Comp_WPA_per_Att = Comp_WPA / Pass_Attempts,
Comp_WPA_per_Comp = Comp_WPA / Completions,
Comp_WPA_per_Drive = Comp_WPA / Pass_Drives,
WPA_Comp_Perc = Comp_WPA / sum(abs(WPA), na.rm = TRUE),
Pass_Clutch_EPA = sum(EPA*abs(WPA), na.rm = TRUE),
Pass_Clutch_EPA_per_Att = Pass_Clutch_EPA / Pass_Attempts,
Pass_Clutch_EPA_per_Comp = Pass_Clutch_EPA / Completions,
Pass_Clutch_EPA_per_Drives = Pass_Clutch_EPA / Pass_Drives,
Total_airEPA = sum(Reception*airEPA,na.rm=TRUE),
airEPA_per_Att = Total_airEPA / Pass_Attempts,
airEPA_per_Comp = Total_airEPA / Completions,
airEPA_per_Drive = Total_airEPA / Pass_Drives,
air_Success_Rate = length(which(airEPA>0)) / Pass_Attempts,
air_Comp_Success_Rate = length(which((Reception*airEPA)>0)) / Pass_Attempts,
Total_airWPA = sum(Reception*airWPA,na.rm=TRUE),
airWPA_per_Att = Total_airWPA / Pass_Attempts,
airWPA_per_Comp = Total_airWPA / Completions,
airWPA_per_Drive = Total_airWPA / Pass_Drives,
air_Win_Rate = length(which(airWPA>0)) / Pass_Attempts,
air_Comp_Win_Rate = length(which((Reception*airWPA)>0)) / Pass_Attempts,
Total_yacEPA = sum(Reception*yacEPA,na.rm=TRUE),
yacEPA_per_Att = Total_yacEPA / Pass_Attempts,
yacEPA_per_Comp = Total_yacEPA / Completions,
yacEPA_per_Drive = Total_yacEPA / Pass_Drives,
yac_Success_Rate = length(which((Reception*yacEPA)>0)) / Pass_Attempts,
Total_yacWPA = sum(Reception*yacWPA,na.rm=TRUE),
yacWPA_per_Att = Total_yacWPA / Pass_Attempts,
yacWPA_per_Comp = Total_yacWPA / Completions,
yacWPA_per_Drive = Total_yacWPA / Pass_Drives,
yac_Win_Rate = length(which((Reception*yacWPA)>0)) / Pass_Attempts) %>%
return
}
#' Calculate Receiving Statistics for Given Splits
#'
#' @param pass_pbp Data frame of passing plays.
#' @param splits String vector of variables to split on, default
#' is "Receiver_ID_Name" (splits are levels to calculate the
#' statistics for).
#' @return Data frame of receiving statistics evaluated at the
#' given level of splits.
#' @examples
#' # Calculate receiving statistics :
#' model_data_list <- calculate_receiving_splits(model_data_list$pass_model_df,
#' "Receiver_ID_Name")
#' @export
calc_receiving_splits <- function(pass_pbp, splits = "Receiver_ID_Name") {
# Turn the splits into the group variables:
split_groups <- lapply(splits, as.symbol)
# Create the Game_Drive column:
pass_pbp %>%
dplyr::mutate(Game_Drive = paste(as.character(GameID),
as.character(Drive), sep = "-")) %>%
# Group by the given splits:
dplyr::group_by_(.dots = split_groups) %>%
# Calculate the passing statistics at the given splits level:
dplyr::summarise(Targets = n(),
Receptions = sum(Reception, na.rm = TRUE),
Rec_Drives = n_distinct(Game_Drive),
Rec_Perc = Receptions / Targets,
Rec_Yards_Gained = sum(Yards.Gained, na.rm = TRUE),
Total_AirYards = sum(AirYards, na.rm = TRUE),
Total_AirYards_Rec = sum(Reception*AirYards, na.rm = TRUE),
Total_YardsAfterCatch = sum(Reception*YardsAfterCatch, na.rm = TRUE),
Yards_per_Tgt = Rec_Yards_Gained / Targets,
Yards_per_Rec = Rec_Yards_Gained / Receptions,
Yards_per_Drive = Rec_Yards_Gained / Rec_Drives,
AirYards_per_Tgt = Total_AirYards / Targets,
AirYards_per_Rec = Total_AirYards / Receptions,
AirYards_per_Drive = Total_AirYards / Rec_Drives,
AirYards_Rec_per_Tgt = Total_AirYards_Rec / Targets,
AirYards_Rec_per_Rec = Total_AirYards_Rec / Receptions,
AirYards_Rec_per_Drive = Total_AirYards_Rec / Rec_Drives,
YardsAfterCatch_per_Tgt = Total_YardsAfterCatch / Targets,
YardsAfterCatch_per_Rec = Total_YardsAfterCatch / Receptions,
YardsAfterCatch_per_Drive = Total_YardsAfterCatch / Rec_Drives,
RACR = Rec_Yards_Gained / Total_AirYards,
Rec_INTs = sum(InterceptionThrown, na.rm = TRUE),
Rec_TDs = sum(Touchdown, na.rm = TRUE),
Air_TDs = sum(as.numeric(YardsAfterCatch == 0) * Touchdown, na.rm = TRUE),
Air_TD_Rate = Air_TDs / Rec_TDs,
TD_to_INT = Rec_TDs / Rec_INTs,
TD_per_Tgt = Rec_TDs / Targets,
Air_TD_per_Tgt = Air_TDs / Targets,
TD_per_Rec = Rec_TDs / Receptions,
Air_TD_per_Rec = Air_TDs / Receptions,
TD_per_Drive = Rec_TDs / Rec_Drives,
Air_TD_per_Drive = Air_TDs / Rec_Drives,
Rec_INTs_per_Tgt = Rec_INTs / Targets,
Rec_INTs_per_Rec = Rec_INTs / Receptions,
Rec_INTs_per_Drive = Rec_INTs / Rec_Drives,
Rec_EPA = sum(EPA, na.rm = TRUE),
Rec_EPA_per_Tgt = Rec_EPA / Targets,
Rec_EPA_per_Rec = Rec_EPA / Receptions,
Rec_EPA_per_Drive = Rec_EPA / Rec_Drives,
Rec_Success_Rate = length(which(EPA > 0)) / Targets,
Caught_EPA = sum(Reception * EPA, na.rm = TRUE),
Caught_EPA_per_Tgt = Caught_EPA / Targets,
Caught_EPA_per_Rec = Caught_EPA / Receptions,
Caught_EPA_per_Drive = Caught_EPA / Rec_Drives,
EPA_Rec_Perc = Caught_EPA / sum(abs(EPA), na.rm = TRUE),
Rec_WPA = sum(WPA, na.rm = TRUE),
Rec_WPA_per_Tgt = Rec_WPA / Targets,
Rec_WPA_per_Rec = Rec_WPA / Receptions,
Rec_WPA_per_Drive = Rec_WPA / Rec_Drives,
Rec_Win_Rate = length(which(WPA > 0)) / Targets,
Caught_WPA = sum(Reception * WPA, na.rm = TRUE),
Caught_WPA_per_Tgt = Caught_WPA / Targets,
Caught_WPA_per_Rec = Caught_WPA / Receptions,
Caught_WPA_per_Drive = Caught_WPA / Rec_Drives,
WPA_Rec_Perc = Caught_WPA / sum(abs(WPA), na.rm = TRUE),
Rec_Clutch_EPA = sum(EPA*abs(WPA), na.rm = TRUE),
Rec_Clutch_EPA_per_Tgt = Rec_Clutch_EPA / Targets,
Rec_Clutch_EPA_per_Rec = Rec_Clutch_EPA / Receptions,
Rec_Clutch_EPA_per_Drives = Rec_Clutch_EPA / Rec_Drives,
Total_airEPA = sum(Reception*airEPA,na.rm=TRUE),
airEPA_per_Tgt = Total_airEPA / Targets,
airEPA_per_Rec = Total_airEPA / Receptions,
airEPA_per_Drive = Total_airEPA / Rec_Drives,
air_Success_Rate = length(which(airEPA>0)) / Targets,
air_Rec_Success_Rate = length(which((Reception*airEPA)>0)) / Targets,
Total_airWPA = sum(Reception*airWPA,na.rm=TRUE),
airWPA_per_Tgt = Total_airWPA / Targets,
airWPA_per_Rec = Total_airWPA / Receptions,
airWPA_per_Drive = Total_airWPA / Rec_Drives,
air_Win_Rate = length(which(airWPA>0)) / Targets,
air_Rec_Win_Rate = length(which((Reception*airWPA)>0)) / Targets,
Total_yacEPA = sum(Reception*yacEPA,na.rm=TRUE),
yacEPA_per_Tgt = Total_yacEPA / Targets,
yacEPA_per_Rec = Total_yacEPA / Receptions,
yacEPA_per_Drive = Total_yacEPA / Rec_Drives,
yac_Success_Rate = length(which((Reception*yacEPA)>0)) / Targets,
Total_yacWPA = sum(Reception*yacWPA,na.rm=TRUE),
yacWPA_per_Tgt = Total_yacWPA / Targets,
yacWPA_per_Rec = Total_yacWPA / Receptions,
yacWPA_per_Drive = Total_yacWPA / Rec_Drives,
yac_Win_Rate = length(which((Reception*yacWPA)>0)) / Targets) %>%
return
}
#' Calculate Rushing Statistics for Given Splits
#'
#' @param rush_pbp Data frame of rushing plays.
#' @param splits String vector of variables to split on, default
#' is "Rusher_ID_Name" (splits are levels to calculate the
#' statistics for).
#' @return Data frame of rushing statistics evaluated at the
#' given level of splits.
#' @examples
#' # Calculate receiving statistics :
#' model_data_list <- calculate_rushing_splits(model_data_list$pass_model_df,
#' "Rusher_ID_Name")
#' @export
calculate_rushing_splits <- function(rush_pbp, splits = "Rusher_ID_Name") {
# Turn the splits into the group variables:
split_groups <- lapply(splits, as.symbol)
# Create the Game_Drive column:
rush_pbp %>%
dplyr::mutate(Game_Drive = paste(as.character(GameID),
as.character(Drive), sep = "-")) %>%
# Group by the given splits:
dplyr::group_by_(.dots = split_groups) %>%
# Calculate the passing statistics at the given splits level:
dplyr::summarise(Rushes = n(),
Sacks = sum(Sack, na.rm = TRUE),
Rush_Drives = n_distinct(Game_Drive),
Rush_Yards_Gained = sum(Yards.Gained, na.rm = TRUE),
Sack_Yards_Lost = sum(Sack * Yards.Gained, na.rm = TRUE),
Rush_Yards_per_Rush = Rush_Yards_Gained / Rushes,
Rush_Yards_per_Drive = Rush_Yards_Gained / Rush_Drives,
Rush_TDs = sum(Touchdown, na.rm = TRUE),
Rush_TD_per_Rush = Rush_TDs / Rushes,
Rush_TD_per_Drive = Rush_TDs / Rush_Drives,
Rush_EPA = sum(EPA, na.rm = TRUE),
Rush_EPA_per_Rush = Rush_EPA / Rushes,
Rush_EPA_per_Drive = Rush_EPA / Rush_Drives,
Rush_Success_Rate = length(which(EPA>0)) / Rushes,
Pos_EPA_Ratio = sum(as.numeric(EPA > 0) * EPA, na.rm = TRUE) / sum(abs(EPA), na.rm = TRUE),
Rush_WPA = sum(WPA, na.rm = TRUE),
Rush_WPA_per_Rush = Rush_WPA / Rushes,
Rush_WPA_per_Drive = Rush_WPA / Rush_Drives,
Rush_Win_Rate = length(which(WPA>0)) / Rushes,
Pos_WPA_Ratio = sum(as.numeric(WPA > 0) * WPA, na.rm = TRUE) / sum(abs(WPA), na.rm = TRUE),
Fumbles = sum(Fumble,na.rm = TRUE),
Fumbles_per_Rush = Fumbles / Rushes,
Fumbles_per_Drive = Fumbles / Rush_Drives,
Rush_Clutch_EPA = sum(EPA*abs(WPA),na.rm=TRUE),
Rush_Clutch_EPA_per_Rush = Rush_Clutch_EPA / Rushes,
Rush_Clutch_EPA_per_Drive = Rush_Clutch_EPA / Rush_Drives) %>%
return
}
ryurko/nflWAR documentation built on May 9, 2019, 7:37 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#' Join Various Statistics to the Position Tables
#'
#' @param model_data_list List that must have the following:
#' \itemize{
#' \item{"pass_model_df"} - Passing play-by-play data
#' \item{"rush_model_df"} - Rushing play-by-play data
#' \item{"QB_table"} - Table of QBs
#' \item{"RB_table"} - Table of RBs
#' \item{"WR_table"} - Table of WRs
#' \item{"TE_table"} - Table of TEs
#' }
#' @return The input position tables with all the different
#' types of statistics included.
#' @examples
#' # Calculate the various types of statistics for each position:
#' model_data_list <- model_data_list %>%
#' join_position_statistics(model_data_list)
#' @export
join_position_statistics <- function(model_data_list) {
# Use the split functions to calculate the different
# types of statistics:
passing_stats <- calc_passing_splits(model_data_list$pass_model_df)
receiving_stats <- calc_receiving_splits(model_data_list$pass_model_df)
rushing_stats <- calculate_rushing_splits(model_data_list$rush_model_df)
# Drop the redundant columns in each dataset:
passing_stats <- passing_stats %>%
dplyr::select(-Pass_Attempts)
receiving_stats <- receiving_stats %>%
dplyr::select(-Targets)
rushing_stats <- rushing_stats %>%
dplyr::select(-Rushes, -Sacks)
# Now join the stats to the position tables, and return:
model_data_list$QB_table <- model_data_list$QB_table %>%
dplyr::left_join(passing_stats,
by = c("Player_ID_Name" = "Passer_ID_Name")) %>%
dplyr::left_join(rushing_stats,
by = c("Player_ID_Name" = "Rusher_ID_Name")) %>%
dplyr::mutate(AdjNetYardsAtt = (Pass_Yards_Gained + 20 * Pass_TDs - 45 * INTs + Sack_Yards_Lost) / (Pass_Attempts + Sacks))
model_data_list$QB_table[is.na(model_data_list$QB_table)] <- 0
# Expression for the non-QB tables:
join_nonQB_stats <- . %>%
dplyr::left_join(receiving_stats,
by = c("Player_ID_Name" = "Receiver_ID_Name")) %>%
dplyr::left_join(rushing_stats,
by = c("Player_ID_Name" = "Rusher_ID_Name"))
model_data_list$WR_table <- model_data_list$WR_table %>% join_nonQB_stats
model_data_list$WR_table[is.na(model_data_list$WR_table)] <- 0
model_data_list$RB_table <- model_data_list$RB_table %>% join_nonQB_stats
model_data_list$RB_table[is.na(model_data_list$RB_table)] <- 0
model_data_list$TE_table <- model_data_list$TE_table %>% join_nonQB_stats
model_data_list$TE_table[is.na(model_data_list$TE_table)] <- 0
return(model_data_list)
}
#' Calculate Passing Statistics for Given Splits
#'
#' @param pass_pbp Data frame of passing plays.
#' @param splits String vector of variables to split on, default
#' is "Passer_ID_Name" (splits are levels to calculate the
#' statistics for).
#' @return Data frame of passing statistics evaluated at the
#' given level of splits.
#' @examples
#' # Calculate passing statistics :
#' model_data_list <- calculate_passing_splits(model_data_list$pass_model_df,
#' "Passer_ID_Name")
#' @export
calc_passing_splits <- function(pass_pbp, splits = "Passer_ID_Name") {
# Turn the splits into the group variables:
split_groups <- lapply(splits, as.symbol)
# Create the Game_Drive column:
pass_pbp %>%
dplyr::mutate(Game_Drive = paste(as.character(GameID),
as.character(Drive), sep = "-")) %>%
# Group by the given splits:
dplyr::group_by_(.dots = split_groups) %>%
# Calculate the passing statistics at the given splits level:
dplyr::summarise(Pass_Attempts = n(),
Completions = sum(Reception, na.rm = TRUE),
Pass_Drives = n_distinct(Game_Drive),
Comp_Perc = Completions / Pass_Attempts,
Pass_Yards_Gained = sum(Yards.Gained, na.rm = TRUE),
Total_AirYards = sum(AirYards, na.rm = TRUE),
Total_AirYards_Comp = sum(Reception*AirYards, na.rm = TRUE),
Total_YardsAfterCatch = sum(Reception*YardsAfterCatch, na.rm = TRUE),
Yards_per_Att = Pass_Yards_Gained / Pass_Attempts,
Yards_per_Comp = Pass_Yards_Gained / Completions,
Yards_per_Drive = Pass_Yards_Gained / Pass_Drives,
AirYards_per_Att = Total_AirYards / Pass_Attempts,
AirYards_per_Comp = Total_AirYards / Completions,
AirYards_per_Drive = Total_AirYards / Pass_Drives,
AirYards_Comp_per_Att = Total_AirYards_Comp / Pass_Attempts,
AirYards_Comp_per_Comp = Total_AirYards_Comp / Completions,
AirYards_Comp_per_Drive = Total_AirYards_Comp / Pass_Drives,
YardsAfterCatch_per_Att = Total_YardsAfterCatch / Pass_Attempts,
YardsAfterCatch_per_Comp = Total_YardsAfterCatch / Completions,
YardsAfterCatch_per_Drive = Total_YardsAfterCatch / Pass_Drives,
PACR = Pass_Yards_Gained / Total_AirYards,
Times_Hit = sum(QBHit, na.rm = TRUE),
Times_Hit_per_Att = Times_Hit / Pass_Attempts,
Times_Hit_per_Comp = Times_Hit / Completions,
Times_Hit_per_Drive = Times_Hit / Pass_Drives,
INTs = sum(InterceptionThrown, na.rm = TRUE),
Pass_TDs = sum(Touchdown, na.rm = TRUE),
Air_TDs = sum(as.numeric(YardsAfterCatch == 0) * Touchdown, na.rm = TRUE),
Air_TD_Rate = Air_TDs / Pass_TDs,
TD_to_INT = Pass_TDs / INTs,
TD_per_Att = Air_TDs / Pass_Attempts,
Air_TD_per_Att = Air_TDs / Pass_Attempts,
TD_per_Comp = Pass_TDs / Completions,
Air_TD_per_Comp = Air_TDs / Completions,
TD_per_Drive = Pass_TDs / Pass_Drives,
Air_TD_per_Drive = Air_TDs / Pass_Drives,
INTs_per_Att = INTs / Pass_Attempts,
Passer_Rating = ((((Comp_Perc - .3)*5) + ((Yards_per_Att - 3)*.25) + (TD_per_Att*20) + (2.375 - (INTs_per_Att * 25))) / 6) * 100,
INTs_per_Comp = INTs / Completions,
INTs_per_Drive = INTs / Pass_Drives,
Pass_EPA = sum(EPA, na.rm = TRUE),
Pass_EPA_per_Att = Pass_EPA / Pass_Attempts,
Pass_EPA_per_Comp = Pass_EPA / Completions,
Pass_EPA_per_Drive = Pass_EPA / Pass_Drives,
Pass_Success_Rate = length(which(EPA > 0)) / Pass_Attempts,
Comp_EPA = sum(Reception * EPA, na.rm = TRUE),
Comp_EPA_per_Att = Comp_EPA / Pass_Attempts,
Comp_EPA_per_Comp = Comp_EPA / Completions,
Comp_EPA_per_Drive = Comp_EPA / Pass_Drives,
EPA_Comp_Perc = Comp_EPA / sum(abs(EPA), na.rm = TRUE),
Pass_WPA = sum(WPA, na.rm = TRUE),
Pass_WPA_per_Att = Pass_WPA / Pass_Attempts,
Pass_WPA_per_Comp = Pass_WPA / Completions,
Pass_WPA_per_Drive = Pass_WPA / Pass_Drives,
Pass_Win_Rate = length(which(WPA > 0)) / Pass_Attempts,
Comp_WPA = sum(Reception * WPA, na.rm = TRUE),
Comp_WPA_per_Att = Comp_WPA / Pass_Attempts,
Comp_WPA_per_Comp = Comp_WPA / Completions,
Comp_WPA_per_Drive = Comp_WPA / Pass_Drives,
WPA_Comp_Perc = Comp_WPA / sum(abs(WPA), na.rm = TRUE),
Pass_Clutch_EPA = sum(EPA*abs(WPA), na.rm = TRUE),
Pass_Clutch_EPA_per_Att = Pass_Clutch_EPA / Pass_Attempts,
Pass_Clutch_EPA_per_Comp = Pass_Clutch_EPA / Completions,
Pass_Clutch_EPA_per_Drives = Pass_Clutch_EPA / Pass_Drives,
Total_airEPA = sum(Reception*airEPA,na.rm=TRUE),
airEPA_per_Att = Total_airEPA / Pass_Attempts,
airEPA_per_Comp = Total_airEPA / Completions,
airEPA_per_Drive = Total_airEPA / Pass_Drives,
air_Success_Rate = length(which(airEPA>0)) / Pass_Attempts,
air_Comp_Success_Rate = length(which((Reception*airEPA)>0)) / Pass_Attempts,
Total_airWPA = sum(Reception*airWPA,na.rm=TRUE),
airWPA_per_Att = Total_airWPA / Pass_Attempts,
airWPA_per_Comp = Total_airWPA / Completions,
airWPA_per_Drive = Total_airWPA / Pass_Drives,
air_Win_Rate = length(which(airWPA>0)) / Pass_Attempts,
air_Comp_Win_Rate = length(which((Reception*airWPA)>0)) / Pass_Attempts,
Total_yacEPA = sum(Reception*yacEPA,na.rm=TRUE),
yacEPA_per_Att = Total_yacEPA / Pass_Attempts,
yacEPA_per_Comp = Total_yacEPA / Completions,
yacEPA_per_Drive = Total_yacEPA / Pass_Drives,
yac_Success_Rate = length(which((Reception*yacEPA)>0)) / Pass_Attempts,
Total_yacWPA = sum(Reception*yacWPA,na.rm=TRUE),
yacWPA_per_Att = Total_yacWPA / Pass_Attempts,
yacWPA_per_Comp = Total_yacWPA / Completions,
yacWPA_per_Drive = Total_yacWPA / Pass_Drives,
yac_Win_Rate = length(which((Reception*yacWPA)>0)) / Pass_Attempts) %>%
return
}
#' Calculate Receiving Statistics for Given Splits
#'
#' @param pass_pbp Data frame of passing plays.
#' @param splits String vector of variables to split on, default
#' is "Receiver_ID_Name" (splits are levels to calculate the
#' statistics for).
#' @return Data frame of receiving statistics evaluated at the
#' given level of splits.
#' @examples
#' # Calculate receiving statistics :
#' model_data_list <- calculate_receiving_splits(model_data_list$pass_model_df,
#' "Receiver_ID_Name")
#' @export
calc_receiving_splits <- function(pass_pbp, splits = "Receiver_ID_Name") {
# Turn the splits into the group variables:
split_groups <- lapply(splits, as.symbol)
# Create the Game_Drive column:
pass_pbp %>%
dplyr::mutate(Game_Drive = paste(as.character(GameID),
as.character(Drive), sep = "-")) %>%
# Group by the given splits:
dplyr::group_by_(.dots = split_groups) %>%
# Calculate the passing statistics at the given splits level:
dplyr::summarise(Targets = n(),
Receptions = sum(Reception, na.rm = TRUE),
Rec_Drives = n_distinct(Game_Drive),
Rec_Perc = Receptions / Targets,
Rec_Yards_Gained = sum(Yards.Gained, na.rm = TRUE),
Total_AirYards = sum(AirYards, na.rm = TRUE),
Total_AirYards_Rec = sum(Reception*AirYards, na.rm = TRUE),
Total_YardsAfterCatch = sum(Reception*YardsAfterCatch, na.rm = TRUE),
Yards_per_Tgt = Rec_Yards_Gained / Targets,
Yards_per_Rec = Rec_Yards_Gained / Receptions,
Yards_per_Drive = Rec_Yards_Gained / Rec_Drives,
AirYards_per_Tgt = Total_AirYards / Targets,
AirYards_per_Rec = Total_AirYards / Receptions,
AirYards_per_Drive = Total_AirYards / Rec_Drives,
AirYards_Rec_per_Tgt = Total_AirYards_Rec / Targets,
AirYards_Rec_per_Rec = Total_AirYards_Rec / Receptions,
AirYards_Rec_per_Drive = Total_AirYards_Rec / Rec_Drives,
YardsAfterCatch_per_Tgt = Total_YardsAfterCatch / Targets,
YardsAfterCatch_per_Rec = Total_YardsAfterCatch / Receptions,
YardsAfterCatch_per_Drive = Total_YardsAfterCatch / Rec_Drives,
RACR = Rec_Yards_Gained / Total_AirYards,
Rec_INTs = sum(InterceptionThrown, na.rm = TRUE),
Rec_TDs = sum(Touchdown, na.rm = TRUE),
Air_TDs = sum(as.numeric(YardsAfterCatch == 0) * Touchdown, na.rm = TRUE),
Air_TD_Rate = Air_TDs / Rec_TDs,
TD_to_INT = Rec_TDs / Rec_INTs,
TD_per_Tgt = Rec_TDs / Targets,
Air_TD_per_Tgt = Air_TDs / Targets,
TD_per_Rec = Rec_TDs / Receptions,
Air_TD_per_Rec = Air_TDs / Receptions,
TD_per_Drive = Rec_TDs / Rec_Drives,
Air_TD_per_Drive = Air_TDs / Rec_Drives,
Rec_INTs_per_Tgt = Rec_INTs / Targets,
Rec_INTs_per_Rec = Rec_INTs / Receptions,
Rec_INTs_per_Drive = Rec_INTs / Rec_Drives,
Rec_EPA = sum(EPA, na.rm = TRUE),
Rec_EPA_per_Tgt = Rec_EPA / Targets,
Rec_EPA_per_Rec = Rec_EPA / Receptions,
Rec_EPA_per_Drive = Rec_EPA / Rec_Drives,
Rec_Success_Rate = length(which(EPA > 0)) / Targets,
Caught_EPA = sum(Reception * EPA, na.rm = TRUE),
Caught_EPA_per_Tgt = Caught_EPA / Targets,
Caught_EPA_per_Rec = Caught_EPA / Receptions,
Caught_EPA_per_Drive = Caught_EPA / Rec_Drives,
EPA_Rec_Perc = Caught_EPA / sum(abs(EPA), na.rm = TRUE),
Rec_WPA = sum(WPA, na.rm = TRUE),
Rec_WPA_per_Tgt = Rec_WPA / Targets,
Rec_WPA_per_Rec = Rec_WPA / Receptions,
Rec_WPA_per_Drive = Rec_WPA / Rec_Drives,
Rec_Win_Rate = length(which(WPA > 0)) / Targets,
Caught_WPA = sum(Reception * WPA, na.rm = TRUE),
Caught_WPA_per_Tgt = Caught_WPA / Targets,
Caught_WPA_per_Rec = Caught_WPA / Receptions,
Caught_WPA_per_Drive = Caught_WPA / Rec_Drives,
WPA_Rec_Perc = Caught_WPA / sum(abs(WPA), na.rm = TRUE),
Rec_Clutch_EPA = sum(EPA*abs(WPA), na.rm = TRUE),
Rec_Clutch_EPA_per_Tgt = Rec_Clutch_EPA / Targets,
Rec_Clutch_EPA_per_Rec = Rec_Clutch_EPA / Receptions,
Rec_Clutch_EPA_per_Drives = Rec_Clutch_EPA / Rec_Drives,
Total_airEPA = sum(Reception*airEPA,na.rm=TRUE),
airEPA_per_Tgt = Total_airEPA / Targets,
airEPA_per_Rec = Total_airEPA / Receptions,
airEPA_per_Drive = Total_airEPA / Rec_Drives,
air_Success_Rate = length(which(airEPA>0)) / Targets,
air_Rec_Success_Rate = length(which((Reception*airEPA)>0)) / Targets,
Total_airWPA = sum(Reception*airWPA,na.rm=TRUE),
airWPA_per_Tgt = Total_airWPA / Targets,
airWPA_per_Rec = Total_airWPA / Receptions,
airWPA_per_Drive = Total_airWPA / Rec_Drives,
air_Win_Rate = length(which(airWPA>0)) / Targets,
air_Rec_Win_Rate = length(which((Reception*airWPA)>0)) / Targets,
Total_yacEPA = sum(Reception*yacEPA,na.rm=TRUE),
yacEPA_per_Tgt = Total_yacEPA / Targets,
yacEPA_per_Rec = Total_yacEPA / Receptions,
yacEPA_per_Drive = Total_yacEPA / Rec_Drives,
yac_Success_Rate = length(which((Reception*yacEPA)>0)) / Targets,
Total_yacWPA = sum(Reception*yacWPA,na.rm=TRUE),
yacWPA_per_Tgt = Total_yacWPA / Targets,
yacWPA_per_Rec = Total_yacWPA / Receptions,
yacWPA_per_Drive = Total_yacWPA / Rec_Drives,
yac_Win_Rate = length(which((Reception*yacWPA)>0)) / Targets) %>%
return
}
#' Calculate Rushing Statistics for Given Splits
#'
#' @param rush_pbp Data frame of rushing plays.
#' @param splits String vector of variables to split on, default
#' is "Rusher_ID_Name" (splits are levels to calculate the
#' statistics for).
#' @return Data frame of rushing statistics evaluated at the
#' given level of splits.
#' @examples
#' # Calculate receiving statistics :
#' model_data_list <- calculate_rushing_splits(model_data_list$pass_model_df,
#' "Rusher_ID_Name")
#' @export
calculate_rushing_splits <- function(rush_pbp, splits = "Rusher_ID_Name") {
# Turn the splits into the group variables:
split_groups <- lapply(splits, as.symbol)
# Create the Game_Drive column:
rush_pbp %>%
dplyr::mutate(Game_Drive = paste(as.character(GameID),
as.character(Drive), sep = "-")) %>%
# Group by the given splits:
dplyr::group_by_(.dots = split_groups) %>%
# Calculate the passing statistics at the given splits level:
dplyr::summarise(Rushes = n(),
Sacks = sum(Sack, na.rm = TRUE),
Rush_Drives = n_distinct(Game_Drive),
Rush_Yards_Gained = sum(Yards.Gained, na.rm = TRUE),
Sack_Yards_Lost = sum(Sack * Yards.Gained, na.rm = TRUE),
Rush_Yards_per_Rush = Rush_Yards_Gained / Rushes,
Rush_Yards_per_Drive = Rush_Yards_Gained / Rush_Drives,
Rush_TDs = sum(Touchdown, na.rm = TRUE),
Rush_TD_per_Rush = Rush_TDs / Rushes,
Rush_TD_per_Drive = Rush_TDs / Rush_Drives,
Rush_EPA = sum(EPA, na.rm = TRUE),
Rush_EPA_per_Rush = Rush_EPA / Rushes,
Rush_EPA_per_Drive = Rush_EPA / Rush_Drives,
Rush_Success_Rate = length(which(EPA>0)) / Rushes,
Pos_EPA_Ratio = sum(as.numeric(EPA > 0) * EPA, na.rm = TRUE) / sum(abs(EPA), na.rm = TRUE),
Rush_WPA = sum(WPA, na.rm = TRUE),
Rush_WPA_per_Rush = Rush_WPA / Rushes,
Rush_WPA_per_Drive = Rush_WPA / Rush_Drives,
Rush_Win_Rate = length(which(WPA>0)) / Rushes,
Pos_WPA_Ratio = sum(as.numeric(WPA > 0) * WPA, na.rm = TRUE) / sum(abs(WPA), na.rm = TRUE),
Fumbles = sum(Fumble,na.rm = TRUE),
Fumbles_per_Rush = Fumbles / Rushes,
Fumbles_per_Drive = Fumbles / Rush_Drives,
Rush_Clutch_EPA = sum(EPA*abs(WPA),na.rm=TRUE),
Rush_Clutch_EPA_per_Rush = Rush_Clutch_EPA / Rushes,
Rush_Clutch_EPA_per_Drive = Rush_Clutch_EPA / Rush_Drives) %>%
return
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.