ESPN Fantasy Football 2020 Part 1: Accessing the API

Extract one player’s information

number_of_teams <- length(ESPNFromJSON$teams$id)
team_ids <- ESPNFromJSON$teams$id

player_extract <- function(team_number = 1, player_number = 1, per_id = per_id, ESPNFromJSON = ESPNFromJSON){
  player_week <-
    tibble(
      team = str_c(ESPNFromJSON$teams$location[team_number]," ",ESPNFromJSON$teams$nickname[team_number]),
      teamId = ESPNFromJSON$teams$id[team_number],
      fullName = ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$fullName[player_number],
      appliedTotal = ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$stats[[player_number]]$appliedTotal,
      seasonId = ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$stats[[player_number]]$seasonId,
      scoringPeriodId = ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$stats[[player_number]]$scoringPeriodId,
      statsplitTypeId = ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$stats[[player_number]]$statSplitTypeId,
      externalId = ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$stats[[player_number]]$externalId,
      lineupSlot_id = ESPNFromJSON$teams$roster$entries[[team_number]]$lineupSlotId[player_number],
      eligibleSlots = list(ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$eligibleSlots[[player_number]])
    ) %>%
    filter(seasonId==2020) %>%
    filter(scoringPeriodId != 0) %>%
    filter(scoringPeriodId == per_id)
  return(player_week)
}

player_extract(ESPNFromJSON = ESPNFromJSON,team_number = 1,player_number = 1,per_id = per_id) %>% gt()

Get all players from all teams

To extract the players for each team, we first need to know how many players are on each team each week. One might think this is consistent, but with an IR slot, this may change from week to week.

First we create a function to determine how many roster spots are on a given team.

get_roster_slots <- function(team_number=1){
  return(tibble(team_number = team_number, player_slot = 1:length(ESPNFromJSON$teams$roster$entries[[team_number]]$playerPoolEntry$player$stats)))
}

Then we map that over all teams.

team_player_slots <- purrr::map_dfr(1:number_of_teams,~get_roster_slots(team_number = .x))

team_player_slots

## # A tibble: 182 x 2
##    team_number player_slot
##          <int>       <int>
##  1           1           1
##  2           1           2
##  3           1           3
##  4           1           4
##  5           1           5
##  6           1           6
##  7           1           7
##  8           1           8
##  9           1           9
## 10           1          10
## # ... with 172 more rows

Now, we map the team_number and player_slots data to the player_extract() function to get every player from every team.

team_list <- 
  purrr::map2_dfr(
    team_player_slots$team_number,
    team_player_slots$player_slot, 
    ~player_extract(
      ESPNFromJSON = ESPNFromJSON,
      team_number = .x,
      player_number = .y, 
      per_id = per_id
      )
    )

team_list %>% 
  group_by(teamId) %>% 
  slice_head(n = 1) %>% 
  ungroup() %>% 
  gt() %>% 
  tab_header("First player on each team")

To get some really good information, we need to join this to the teams and their weekly schedule.

So lets extract the schedule….

schedule <-
  tibble(
    home = ESPNFromJSON$schedule$away$teamId,
    away = ESPNFromJSON$schedule$home$teamId,
    scoringPeriodId = ESPNFromJSON$schedule$matchupPeriodId,
    gameId = ESPNFromJSON$schedule$id
  ) %>%
  pivot_longer(cols = c(home,away), values_to = "teamId")

Then join the players and schedule.

team_list <-
team_list %>% 
  left_join(schedule) %>%
  mutate(points_type = if_else(str_length(externalId) > 6, "actual", "projected")) %>%
  relocate(team:appliedTotal, points_type)

team_list %>% 
  group_by(teamId) %>% 
  slice_head(n = 1) %>% 
  ungroup() %>% 
  gt() %>% 
  tab_header("First player on each team")

Now, we can really start analyzing our league! But that’s for the next post!