The data set All Year Olympic Dataset (with 2020 Tokyo Olympics).csv is a reduced (some variables are missing) extension of the athletes_events.csv till the 2020 Tokyo Olympics. I decided to use this file.
> setwd("C:/Users/vlado/DL/data/kaggle/Olympic/2020")
> source("https://raw.githubusercontent.com/bavla/Rnet/master/R/Pajek.R")
> source("https://raw.githubusercontent.com/bavla/ibm3m/master/multiway/MWnets.R")
> OL <- read.csv("All Year Olympic Dataset (with 2020 Tokyo Olympics).csv")
> dim(OL)
[1] 275643 10
> str(OL)
'data.frame': 275643 obs. of 10 variables:
$ X : int 0 1 2 3 4 5 6 7 8 9 ...
$ Name : chr "A Dijiang" "A Lamusi" "Gunnar Nielsen Aaby" "Edgar Lindenau Aabye" ...
$ Sex : chr "M" "M" "M" "M" ...
$ Age : num 24 23 24 34 21 21 25 25 27 27 ...
$ Team : chr "China" "China" "Denmark" "Denmark/Sweden" ...
$ Games : chr "1992 Summer" "2012 Summer" "1920 Summer" "1900 Summer" ...
$ Year : int 1992 2012 1920 1900 1988 1988 1992 1992 1994 1994 ...
$ Season: chr "Summer" "Summer" "Summer" "Summer" ...
$ Sport : chr "Basketball" "Judo" "Football" "Tug-Of-War" ...
$ Medal : int 0 0 0 3 0 0 0 0 0 0 ...
> OM <- OL[OL$Medal>0,]
> str(OM)
'data.frame': 41404 obs. of 10 variables:
$ X : int 3 37 38 40 41 42 44 48 50 60 ...
$ Name : chr "Edgar Lindenau Aabye" "Arvo Ossian Aaltonen" "Arvo Ossian Aaltonen" "Juhamatti Tapio Aaltonen" ...
$ Sex : chr "M" "M" "M" "M" ...
$ Age : num 34 30 30 28 28 28 28 28 32 20 ...
$ Team : chr "Denmark/Sweden" "Finland" "Finland" "Finland" ...
$ Games : chr "1900 Summer" "1920 Summer" "1920 Summer" "2014 Winter" ...
$ Year : int 1900 1920 1920 2014 1948 1948 1948 1948 1952 1992 ...
$ Season: chr "Summer" "Summer" "Summer" "Winter" ...
$ Sport : chr "Tug-Of-War" "Swimming" "Swimming" "Ice Hockey" ...
$ Medal : int 3 1 1 1 1 3 3 3 1 3 ...
> write.csv(OM,file="medals2020.csv",fileEncoding="UTF-8")
> OS <- OM[OM$Season=="Summer",]
> str(OS)
'data.frame': 35709 obs. of 10 variables:
$ X : int 3 37 38 41 42 44 48 50 79 86 ...
$ Name : chr "Edgar Lindenau Aabye" "Arvo Ossian Aaltonen" "Arvo Ossian Aaltonen" "Paavo Johannes Aaltonen" ...
$ Sex : chr "M" "M" "M" "M" ...
$ Age : num 34 30 30 28 28 28 28 32 27 24 ...
$ Team : chr "Denmark/Sweden" "Finland" "Finland" "Finland" ...
$ Games : chr "1900 Summer" "1920 Summer" "1920 Summer" "1948 Summer" ...
$ Year : int 1900 1920 1920 1948 1948 1948 1948 1952 2008 1920 ...
$ Season: chr "Summer" "Summer" "Summer" "Summer" ...
$ Sport : chr "Tug-Of-War" "Swimming" "Swimming" "Gymnastics" ...
$ Medal : int 3 1 1 1 3 3 3 1 3 2 ...
> write.csv(OS,file="Smedals2020.csv",fileEncoding="UTF-8")
> OW <- OM[OM$Season=="Winter",]
> str(OW)
'data.frame': 5695 obs. of 10 variables:
$ X : int 40 60 61 63 64 67 73 76 78 110 ...
$ Name : chr "Juhamatti Tapio Aaltonen" "Kjetil Andr Aamodt" "Kjetil Andr Aamodt" "Kjetil Andr Aamodt" ...
$ Sex : chr "M" "M" "M" "M" ...
$ Age : num 28 20 20 22 22 22 30 30 34 23 ...
$ Team : chr "Finland" "Norway" "Norway" "Norway" ...
$ Games : chr "2014 Winter" "1992 Winter" "1992 Winter" "1994 Winter" ...
$ Year : int 2014 1992 1992 1994 1994 1994 2002 2002 2006 1952 ...
$ Season: chr "Winter" "Winter" "Winter" "Winter" ...
$ Sport : chr "Ice Hockey" "Alpine Skiing" "Alpine Skiing" "Alpine Skiing" ...
$ Medal : int 1 3 1 2 1 2 3 3 3 1 ...
> write.csv(OW,file="Wmedals2020.csv",fileEncoding="UTF-8")
> medal <- c("Bronze","Silver","Gold")
> OS$Medal <- medal[OS$Medal]
> OS$Sex[OS$Sex=="M"] <- "Male"
> OS$Sex[OS$Sex=="F"] <- "Female"
> MT <- DF2MWN(OS,c("Name","Games","Team","Year","Sport","Sex","Medal"),
+ w="Age",network="Olympic20S0",title="Summer Olympic medals till 2020")
> imed <- c(1,3,2)
> MT$links$Medal <- imed[MT$links$Medal]
> MT$nodes$Medal$ID <- c("Bronze","Silver","Gold")
> str(MT)
> str(MT)
List of 6
$ format: chr "MWnets"
$ info :List of 4
..$ network: chr "Olympic20S0"
..$ title : chr "Summer Olympic medals till 2020"
..$ by : chr "DF2MWN"
..$ date : chr "Sat Feb 4 22:02:21 2023"
$ ways :List of 7
..$ Name : chr "Name"
..$ Games: chr "Games"
..$ Team : chr "Team"
..$ Year : chr "Year"
..$ Sport: chr "Sport"
..$ Sex : chr "Sex"
..$ Medal: chr "Medal"
$ nodes :List of 7
..$ Name :'data.frame': 26114 obs. of 1 variable:
.. ..$ ID: chr [1:26114] "A. Albert" "A. Dubois" "A. Joshua \"Josh\" West" "A. Lawry" ...
..$ Games:'data.frame': 30 obs. of 1 variable:
.. ..$ ID: chr [1:30] "1896 Summer" "1900 Summer" "1904 Summer" "1906 Summer" ...
..$ Team :'data.frame': 489 obs. of 1 variable:
.. ..$ ID: chr [1:489] "A North American Team" "Afghanistan" "Algeria" "Ali-Baba II" ...
..$ Year :'data.frame': 30 obs. of 1 variable:
.. ..$ ID: chr [1:30] "1896" "1900" "1904" "1906" ...
..$ Sport:'data.frame': 65 obs. of 1 variable:
.. ..$ ID: chr [1:65] "Aeronautics" "Alpinism" "Archery" "Art Competitions" ...
..$ Sex :'data.frame': 2 obs. of 1 variable:
.. ..$ ID: chr [1:2] "Female" "Male"
..$ Medal:'data.frame': 3 obs. of 1 variable:
.. ..$ ID: chr [1:3] "Bronze" "Silver" "Gold"
$ links :'data.frame': 35709 obs. of 9 variables:
..$ one : num [1:35709] 1 1 1 1 1 1 1 1 1 1 ...
..$ Name : int [1:35709] 5630 2199 2199 18636 18636 18636 18636 18636 19726 807 ...
..$ Games: int [1:35709] 2 7 7 12 12 12 12 13 27 7 ...
..$ Team : int [1:35709] 106 143 143 143 143 143 143 143 316 316 ...
..$ Year : int [1:35709] 2 7 7 12 12 12 12 13 27 7 ...
..$ Sport: int [1:35709] 57 50 50 29 29 29 29 29 30 29 ...
..$ Sex : int [1:35709] 2 2 2 2 2 2 2 2 1 2 ...
..$ Medal: num [1:35709] 3 1 1 1 3 3 3 1 3 2 ...
..$ Age : num [1:35709] 34 30 30 28 28 28 28 32 27 24 ...
$ data : list()
> write(toJSON(MT),"Olympics20S0.json")
> MD <- MT
> L <- MD$links[,!(names(MD$links) %in% c("Name","Year"))]
> Links <- aggregate(L[,c("one","Age")],by=list(L$Games,L$Team,L$Sport,L$Sex,L$Medal),FUN=sum)
> names(Links) <- c("Games","Team","Sport","Sex","Medal","w","Age")
> MD$links <- Links
> ways <- c("Games","Team","Sport","Sex","Medal"); Ways <- as.list(ways); names(Ways) <- ways
> MD$ways <- Ways; N <- MT$nodes
> MD$nodes <- list(N$Games,N$Team,N$Sport,N$Sex,N$Medal)
> names(MD$nodes) <- ways
> MD$info$network <- "Olympic20S"
> str(MD)
List of 6
$ format: chr "MWnets"
$ info :List of 4
..$ network: chr "Olympic20S"
..$ title : chr "Summer Olympic medals till 2020"
..$ by : chr "DF2MWN"
..$ date : chr "Sat Feb 4 22:02:21 2023"
$ ways :List of 5
..$ Games: chr "Games"
..$ Team : chr "Team"
..$ Sport: chr "Sport"
..$ Sex : chr "Sex"
..$ Medal: chr "Medal"
$ nodes :List of 5
..$ Games:'data.frame': 30 obs. of 1 variable:
.. ..$ ID: chr [1:30] "1896 Summer" "1900 Summer" "1904 Summer" "1906 Summer" ...
..$ Team :'data.frame': 489 obs. of 1 variable:
.. ..$ ID: chr [1:489] "A North American Team" "Afghanistan" "Algeria" "Ali-Baba II" ...
..$ Sport:'data.frame': 65 obs. of 1 variable:
.. ..$ ID: chr [1:65] "Aeronautics" "Alpinism" "Archery" "Art Competitions" ...
..$ Sex :'data.frame': 2 obs. of 1 variable:
.. ..$ ID: chr [1:2] "Female" "Male"
..$ Medal:'data.frame': 3 obs. of 1 variable:
.. ..$ ID: chr [1:3] "Bronze" "Silver" "Gold"
$ links :'data.frame': 11392 obs. of 7 variables:
..$ Games: int [1:11392] 26 29 20 27 25 5 26 30 30 28 ...
..$ Team : int [1:11392] 80 80 143 147 163 168 168 168 213 216 ...
..$ Sport: int [1:11392] 3 3 3 3 3 3 3 3 3 3 ...
..$ Sex : int [1:11392] 1 1 1 1 1 1 1 1 1 1 ...
..$ Medal: num [1:11392] 1 1 1 1 1 1 1 1 1 1 ...
..$ w : num [1:11392] 3 3 1 3 3 1 1 1 1 3 ...
..$ Age : num [1:11392] 63 71 27 86 99 40 32 32 24 67 ...
$ data : list()
> write(toJSON(MD),"Olympics20S.json")