I wanted to learn more about pandas, a popular Python data handling library, and to motivate myself, I used the League of Legends dataest provided by Oracle’s Elixir to do some simple data handling. This was written on the 16th of August, 2018.
The data can be found here.
The goal will be to create a function that can display the games a team played, the opponent, and some useful information. The challnge comes from the fact that each row presents data from a single team’s point of view. The data is effectively duplicated again from the opponent’s view. For example, a single row looks like:’
| gameid | date | week | gamelength | ... |But no opponent field. We can get the opponent by joining on the gameid field.
Let’s start by getting all the gameid that relevant to a team, which we will pass as a command line argument. The program will run like this: python3 team_games.py "Team Liquid".
import sys
import pandas as pd
team = sys.argv[1]
df = pd.read_csv("data.csv")
fields = 'gameid date week game team side gamelength result teamkills teamdeaths'.split()
team_games = df[df['team'] == team].drop_duplicates(subset=['gameid'])[fields]
team_game_ids = team_games['gameid']data.csv is the data from Oracle’s Elixir. It is an excel spreadsheet, so I had to open it, then export is as a .csv file, so I could read it using pandas.
We startt by getting any games where the team column is the team we passed in as the first argument. Each game is displayed 10 times, once for each player on each team. We don’t care about individual players for this program, so we drop any duplicated gameid rows using drop_duplicates and passing the gameid as a subset. Next we grab just the gameid column. This will be useful for performing the join later.
Now we will get all rows that were in any of the games played by the team. This will include the opponents:
games = df.loc[df['gameid'].isin(team_game_ids)]
unique_fields = ['gameid']
opponents = games[games['team'] != team].drop_duplicates(subset=unique_fields)[['gameid', 'team']]
opponents.columns = ['gameid', 'opponent']We then proceed to do the reverse logic to the first step: get the rows where the team is not the team we passed in. This time, we only want the gameid and team This time, we only want the gameid and team. We will also rename the columns, since otherwise we will end up with two team columns.
Lastly, we join the data on the gameid column. We will also move the opponent column to be directly next to the team column. Since you can’t have two columns with the same name, I had to rename opponent to the_opponent when reinserting the column at a different location.
result = pd.merge(team_games, opponents, on='gameid')
result.insert(5, 'the_opponent', result['opponent'])
result.drop('opponent', axis=1, inplace=True)
print(result)Running this with python3 team_games.py "Echo Fox" gives me:
gameid week game team side the_opponent gamelength result teamkills teamdeaths
0 1002620109 1.0 2 Echo Fox Red Clutch Gaming 32.430000 1 15.0 15.0
1 1002620050 1.0 1 Echo Fox Blue FlyQuest 23.833333 1 18.0 13.0
2 1002650038 2.0 1 Echo Fox Red OpTic Gaming 38.900000 0 9.0 15.0
3 1002650084 2.0 2 Echo Fox Blue Golden Guardians 34.733333 1 23.0 13.0
4 1002650117 3.0 1 Echo Fox Red Team Liquid 22.300000 0 7.0 28.0
5 1002650174 3.0 2 Echo Fox Red Cloud9 37.366667 1 20.0 18.0
6 1002670206 4.0 1 Echo Fox Blue Counter Logic Gaming 36.483333 0 5.0 25.0
7 1002670249 4.0 2 Echo Fox Red Team SoloMid 31.950000 1 10.0 4.0
8 1002660285 5.0 1 Echo Fox Blue 100 Thieves 41.066667 0 10.0 20.0
9 1002660352 5.0 2 Echo Fox Red Golden Guardians 32.383333 1 10.0 11.0
10 1002690019 6.0 1 Echo Fox Blue FlyQuest 25.916667 1 13.0 3.0
11 1002690064 6.0 2 Echo Fox Red Cloud9 36.083333 0 13.0 13.0
12 1002680110 7.0 1 Echo Fox Red Counter Logic Gaming 40.600000 1 18.0 17.0
13 1002680146 7.0 2 Echo Fox Blue Team SoloMid 43.583333 1 14.0 14.0
14 1002710017 8.0 1 Echo Fox Red 100 Thieves 28.633333 0 4.0 16.0
15 1002710059 8.0 2 Echo Fox Blue OpTic Gaming 36.416667 0 9.0 10.0Great! The next step is to do some analysis, and maybe make some visualisations.