weekly_values_mlm: Generates weekly team-unit estimates

In mattmills49/MELLS: Build CFB points scored prediction model

Description

This function takes in the formatted play by play data and build weekly team estimates to be used in the point prediction models. This function uses multilevel models on play by play data prior to a given week to build the estimates going in to that week.

Usage

weekly_values_mlm(model_runs, model_pass, model_drives)

Arguments

model_runs	the cleaned run plays after the add_epa function
model_pass	the cleaned pass plays after the add_epa function
model_drive	the fixed drives after the fix_drives function

Value

A data frame containing the weekly teem estimates

Examples

years <- 2013:2014
plays <- readin("play", years)
teams <- readin("team", years)
runs <- readin("rush", years)
pass <- readin("pass", years)
games <- readin("game", years)
conf <- readin("conference", years)

epa_model <- expected_points_build(plays[plays$Year != 2014, ], drives[drives$Year != 2014, ])
fixed_games <- fix_games(games)
drives <- fix_drives(fixed_games, drives)
model_plays <- combine_run_pass(runs, pass, fixed_games) %>% remove_garbage %>% fix_fcs(teams, conf) %>% add_epa(epa_model)
model_values <- generate_preseason_mlm(run_plays = model_plays[["run_info"]], pass_plays = model_plays[["pass_info"]]) %>% add_preseason(run_plays = model_plays[["run_info"]], pass_plays = model_plays[["pass_info"]], drives = drives, preseason_vals = .)
weekly_values <- weekly_values_mlm(model_runs = model_values[["run_data"]], model_pass = model_values[["pass_data"]], model_drives = model_values[["drives_data"]])

mattmills49/MELLS documentation built on May 21, 2019, 1:25 p.m.