Football History Time Lapse

Introduction

How did the beautiful game of football propogate throughout the world?

I am a bit ‘Sleepless in Seattle’ on Saturday morning and really excited for the World Cup which is going to kick-off next week on June 14th.

To get in the mood for the upcoming month, I found myself a dataset with all the international football matches ever played since the 1872 friendly between England vs Scotland which was held in Glasgow. Subsequently, I created a geo plot gif which highlights all nations which played any international football with size of circle corresponding to count of matches for the frame i.e. year. The circles become darker as years go by because they are retained over frames to give a cumulative summary. Thereby, giving us a sense of how the game became popular across the world!

Animation

Football has its origins in the 1870s England from where it got spread to the rest of Europe, then in South America during the early 1900s followed by Africa around the 1940s and finally Asia in the 1960s. Bear in mind above animation only accounts for count of matches and not the quality or FIFA ranking.

Interestingly, India did not play its first match till 1938 friendly against Australia at away where it lost 5-3 while Indian players were playing bare feet. The legend it seems is true!

HERE IS THE CODE FOR CREATING ABOVE ANIMATION:

Importing required libraries

library(ggplot2)
library(dplyr)
library(DT)
library(maps)
library(ggthemes)
library(plotly)
library(tibble)
library(lubridate)
library(gganimate)
library(readr)
library(ggmap)
library(animation)

Data Exhibit

Glancing at the dataset

head(df)

date	home_team	away_team	home_score	away_score	tournament	city	country	neutral
1872-11-30	Scotland	England	0	0	Friendly	Glasgow	Scotland	FALSE
1873-03-08	England	Scotland	4	2	Friendly	London	England	FALSE

Code

Some data wrangling to get a data frame with counts of matches for all countries and years

df_home = df %>% select(date, home_team) %>% rename(country = home_team)
df_away = df %>% select(date, away_team) %>% rename(country = away_team)

df_all = rbind(df_home,df_away)
df_all$date = year(df_all$date)
df_all = df_all %>% group_by(date, country) %>% summarise(matches = n()) %>% rename(year = date)

Getting the map coordinates

unique_countries = unique(df_all$country)

map_data <- data.frame(
                      country=unique_countries,
                      lat=1:length(unique_countries),
                      lon=1:length(unique_countries),
                      stringsAsFactors=FALSE
                      )


for (i in 1:nrow(map_data)){
  call <- geocode(as.character(map_data$country[i]))
  map_data$lat[i] <- call$lat
  map_data$lon[i] <- call$lon
}

head(map_data)

Static map

world <- ggplot() +
  borders("world", colour = "gray85", fill = "gray80") +
  theme_map() 

map <- world +
  geom_point(aes(x = lon, y = lat, size = matches),
             data = map_data_all, 
             colour = 'purple', alpha = .5) +
  scale_size_continuous(range = c(1, 8), 
                        breaks = c(250, 500, 750, 1000)) +
  labs(size = 'matches')
map

ggplotly(map)

Introducing ghost points to see the final result for a bit longer

ghost_points_ini <- tibble(
  year = '1872',
  matches = 0, lon = 0, lat = 0)

ghost_points_fin <- tibble(
  year = rep('2017', 15),
  matches = 0, lon = 0, lat = 0)

Animate the map

map <- world +
  geom_point(aes(x = lon, y = lat, size = matches, 
                 frame = year,
                 cumulative = TRUE),
             data = map_data_all %>% filter(year > 2000), colour = 'purple', alpha = .1) +
  geom_point(aes(x = lon, y = lat, size = matches, # this is the init transparent frame
                 frame = year,
                 cumulative = TRUE),
             data = ghost_points_ini, alpha = 0) +
  geom_point(aes(x = lon, y = lat, size = matches, # this is the final transparent frames
                 frame = year,
                 cumulative = TRUE),
             data = ghost_points_fin, alpha = 0) +
  scale_size_continuous(range = c(1, 8), breaks = c(250, 500, 750, 1000)) +
  labs(size = 'Count of Matches')  +
    labs(title = 'Year =') +
    theme(plot.title = element_text(size = 20))

library(gganimate)
ani.options(interval = 0.2)
gganimate(map, ani.width = 1000, ani.height = 750, interval=0.2)