The most difficult part of this project was gathering all the data from my Yahoo fantasy league. In order to gain acess to the data in my fantasy league I needed to acess Yahoo’s API

I created a JSON file with my yahoo consumer key and consumer secret

authf <- "../data/auth.json"
auth_dat <- fromJSON(authf)

To create the auth.JSON file one must get consumer_secret and consumer_key from: https://developer.yahoo.com/apps/YWVpEk5c/

Then create a list with the key and secret, convert the list to JSON string and write that JSON string as a file

x <- list(consumer_secret="MYCONSUMERSECRET",consumer_key="MYCONSUMERKEY)
json_txt <- toJSON(x,auto_unbox=T)
write(json_txt,"../data/auth.json")

Yahoo’s API uses the Oauth protocol. Oauth allows users to share their yahoo informaton without having to give out their Username and password.

endpoint <- oauth_endpoint("get_request_token", "request_auth", "get_token",
                           base_url = "https://api.login.yahoo.com/oauth2")
app <- oauth_app("yahoo",
                 key = auth_dat["consumer_key"],
                 secret = auth_dat["consumer_secret"],
                 redirect_uri = "oob")

token <- oauth2.0_token(endpoint, app, use_oob = TRUE, as_header = TRUE,
                        use_basic_auth = TRUE)

config <-  httr::config(token = token)

We now need to tell Yahoo that we want info from BASEBALL fantasy leagues and no other sport.

ff_base <- "https://fantasysports.yahooapis.com/fantasy/v2"
ff.url <- paste0(ff_base,"/game/mlb?format=json")
game.key.json <- fromJSON(as.character(GET(ff.url, config)))
game.key <- game.key.json$fantasy_content$game["game_key"]`

To get info from your specific league you have to use your league id.

league.id <- "MyLeagueID"
league.key <- paste0(game.key, ".l.", league.id)

Once you have acess to your league you can pull any information from the Yahoo API. The info comes in JSON or XML. JSON can be kind of hard to interpret. So I chose to read it in as xml which is a bit easier to understand.

The first bit of Info I grabbed were the team names.

pull_team_names <- function(league.key,config){

  league_url <- sprintf("https://fantasysports.yahooapis.com/fantasy/v2/league/%s/teams/metadata?format=xml", league.key)
  teams_dat <- GET(league_url,config)
  teams_xml <- read_xml(as.character(teams_dat)) %>% xml_ns_strip()
  all_teams <- xml_find_all(teams_xml,"//team")
  team_df <- data_frame(
      team_key=xml_text(xml_find_all(all_teams,"team_key")),
    team_id=xml_text(xml_find_all(all_teams,"team_id")),
    team_name=xml_text(xml_find_all(all_teams,"name")),
  )
  return(team_df)
}

The xml_find_all function is great. It allows you gather all information under nodes with specific names. The team names were stored in the value “name”

pull_stat_categories<- function(config){
  ff_base <- "https://fantasysports.yahooapis.com/fantasy/v2"
  stat_url=paste0(ff_base,"/game/370/stat_categories/?format=xml")
  stat_categories <- read_xml(as.character(GET(stat_url,config))) %>% xml_ns_strip()
  all_stats <- xml_find_all(stat_categories,"//stat")
  all_names <- all_stats %>% map_df(~data_frame(
    stat_name=xml_text(xml_find_all(.x,"name")),
    stat_id=xml_text(xml_find_all(.x,"stat_id"))))
  return(all_names)
}

This function grabs the names of all the stat categories. There are 90 stat categories, but my league only used 10.

get_team_week_stats <- function(team_key,week,config){
  ff_base <- "https://fantasysports.yahooapis.com/fantasy/v2"
  stat_url=paste0(ff_base,"/team/",team_key,"/stats;type=week;week=",week)
  stat_res<- GET(stat_url,config)
  stat_xml <- read_xml(as.character(stat_res)) %>% xml_ns_strip() %>% xml_find_all("//stat")  %>% map_df(~data_frame(
    stat_id=xml_text(xml_find_all(.x,"stat_id")),
    stat_value=xml_text(xml_find_all(.x,"value")))) %>% mutate(week=week,team=team_key)
  return(stat_xml)
}

The get_team_week_stats function will grab the stats from a pariicular team for a certain week.

Create data frames with all team names, and another with all the stat categories

team_df <- pull_team_names(league.key,config)
stat_categories <- pull_stat_categories(config)

weeks <- data_frame(week=1:24)
weeks <- weeks %>% mutate(c=1)
team_df <- team_df %>% mutate(c=1)
team_df <-inner_join(weeks,team_df,by="c")
all_week_stats <- group_by(team_df,team_key,team_name,week) %>% do(get_team_week_stats(.$team_key,.$week,config))%>%ungroup()

In the last line of code I grab all the stats from every week and every team and put into one data frame

all_week_stats <- inner_join(stat_categories,all_week_stats,by="stat_id")

all_week_stats <- all_week_stats%>%
  mutate(stats=ifelse(Stat_Name=="(Walks + Hits)/ Innings Pitched","WHIP",stat_name))%>%
  select(-team_name,-team,-team_key,-stat_id,-stat_name)

After I have my large data frame of stats, I manipulated it a bit. Unti I was left with a data frame of 4 columns:Stat_Name,stat_value,week,Manager

I then saved the data frame into my computer

write_delim(all_stats_hits_innings,"/Users/noahknoblauch/Baseball/all_stats_hits_innings.txt",delim="\t")

The code to do all the analysis can be found here analysis

---

This R Markdown site was created with workflowr

---