In ntyndall/footballstats: Classify and predict football results

Table Of Contents

1. Redis Data Structures
2. Competition Information
3. Match Information
4. Team Information
5. Player Information
6. League Table
7. Results

Redis Data Structures {#datastructures}

Below is a list of data structures that are available through the redis DB

• Competition Information
• Match Information
• Team Information
• Player Information
• League Table

library(footballstats)
library(rredis)
library(magrittr)
blank <- utils::capture.output(footballstats::redis_con())
first_four <- function(z) {
  z <- lapply(1:4, function(x) z[[x]])
  return(c(z, '...'))
}
f_json <- function(y) {
  return(y %>% jsonlite::toJSON(auto_unbox = TRUE) %>% jsonlite::prettify())
}

Competition Information {#competitioninfo}

competition:set

#competition.set <- "competition:set" %>% 
#  rredis::redisSMembers() %>%
competition.set <- list(
  "1005",
  "1007",
  "1093",
  "1102",
  "..."
) %>% f_json()
print(competition.set)

Match Information {#matchinfo}

csm:{comp_id}:{season}:{match_id}

#match.info <- "csm:1204:2017:2213059" %>% 
#  rredis::redisHGetAll() %>%
match.info <- list(
  id =  "2213059",
  comp_id = "1204",
  formatted_date = "09.12.2017",
  season = "2017/2018",
  week = "16",
  venue = "The John Smith's Stadium (Huddersfield, West Yorkshire)",
  venue_id = "15188",
  venue_city = "Huddersfield, West Yorkshire",
  status = "FT",
  timer = "",
  time = "15:00",
  localteam_id = "9220",
  localteam_name = "Huddersfield Town",
  localteam_score = "2",
  visitorteam_id = "9065",
  visitorteam_name = "Brighton & Hove Albion",
  visitorteam_score = "0",
  ht_score = "[2-0]",
  ft_score = "[2-0]",
  et_score = NA,
  penalty_local = NA,
  penalty_visitor = NA
) %>% f_json()
print(match.info)

c:1204:ready.

print(list("2213173", "2213174", "2213175", "2213176", "...") %>% f_json())

Check to see if the match belongs to the set

#match.in.set <- "c_matchSetInfo:1204" %>% 
#  rredis::redisSMembers() %>%
match.in.set <- list(
  "2212907",
  "2212908",
  "2212909",
  "2212910",
  "..."
) %>% f_json()
print(match.in.set)

Team Information {#teaminformation}

This list is constantly pushed to and lifted from with no replace. It looks like the following when we issue lrange analyseTeams 0 -1.

print(list("9221", "9074", "9348", "9021", "...") %>% f_json())

hgetall ct_basic:1204:9065

basic.info <- list(
  team_id = "9065",
  is_national = "False",
  name = "Brighton & Hove Albion",
  country = "England",
  founded = "1901",
  leagues = "1204,1198,1199",
  venue_name = "The American Express Community Stadium",
  venue_id = "86109",
  venue_surface = "grass",
  venue_address = "Village Way",
  venue_city = "Falmer, East Sussex",
  venue_capacity = "30750",
  coach_name = "C. Hughton",
  coach_id = "136328"
) %>% f_json()
print(basic.info)

Player Information {#playerinformation}

Everytime a match is analysed and the team information stored, all the player IDs involved in the update are pushed to a new list, and are accessible through lrange analysePlayers 0 -1. Through another approach, and another CRON job, the players are analysed one by one.

print(list("453145", "49530", "376371", "15512", "...") %>% f_json())

Player statistics are then accessible through the following key names and all have a similar data structure.

player.info <- list(
  name = "West Bromwich Albion",
  id = "9426",
  league = "Premier League",
  league_id = "1204",
  season = "2017/2018",
  minutes = "180",
  appearances = "2",
  lineups = "2",
  substitute_in = "0",
  substitute_out = "0",
  substitute_on_bench = "0",
  goals = "0",
  yellowcards = "0",
  yellowred = "0",
  redcards = "0"
) 
print(player.info %>% f_json())

League Table {#leaguetable}

Lets have a look at the key structure first

#team.info.keys <- "cwt_l:1204:2017*" %>% 
#  rredis::redisKeys() %>% 
team.info.keys <- list(
  "cwt_l:1204:2017:12:9423",
  "cwt_l:1204:2017:15:9158",
  "cwt_l:1204:2017:16:9053",
  "cwt_l:1204:2017:20:9426",
  "..."
) %>% f_json()
print(team.info.keys)

Then looking at a particular key we can see the structure each one has.

team.week.info <- list(
  TEAM = "Newcastle United",
  PTS = "23",
  GF = "21",
  GD = "-10"
) %>% f_json()
print(team.week.info)

Once a key is analysed it is pushed to here;

league.updated <- list(
  "2243999:17002", 
  "2243932:17072", 
  "2239831:14378", 
  "2224810:10307",
  "..."
) %>% f_json()
print(league.updated)

All these individual hash maps containing a single teams accumulative points for each week means we can build a league table for any of the weeks up until the current date, and is accessible through the following;

positions <- list(
  `9259` = "1",
  `9260` = "2",
  `9249` = "3",
  `9092` = "4",
  `...` = "..."
) %>% f_json()
print(positions)

Results {#results}

Below is a list of the results obtained so far...

ntyndall/footballstats documentation built on Aug. 13, 2019, 1:36 p.m.