CODEWITHSUNDEEP
pythonpandas
HD5450
HD5450

Reputation: 53

Updating win rate over time in pandas

So I have a table that looks like this, and I would like to calculate the win rate after each game played. So for player 1, the win rate will start at 100% since he won the first game, then 50% since he lost the second game, then 33%, then back to 50% again.

PlayerId     Status
1               1    
1               0        
1               0    
1               1

so the final table will look something like this

PlayerId     Status     Win_Rate
1               1          100%
1               0          50%
1               0          33%
1               1          50%

Upvotes: 4

Views: 248

Answers (1)

Randy
Randy

Reputation: 14847

You can .groupby PlayerId and use expanding().mean():

In [12]: df.groupby("PlayerId").expanding().mean()
Out[12]:
            PlayerId    Status
PlayerId
1        0       1.0  1.000000
         1       1.0  0.500000
         2       1.0  0.333333
         3       1.0  0.500000
2        4       2.0  1.000000
         5       2.0  0.500000
         6       2.0  0.333333
         7       2.0  0.500000

You can then drop the index level and assign it as a column if you want it on your original frame:

In [18]: df['Win_Rate'] = df.groupby("PlayerId").expanding().mean()['Status'].droplevel(0)

In [19]: df
Out[19]:
   PlayerId  Status  Win_Rate
0         1       1  1.000000
1         1       0  0.500000
2         1       0  0.333333
3         1       1  0.500000
4         2       1  1.000000
5         2       0  0.500000
6         2       0  0.333333
7         2       1  0.500000

If you want it as a string percentage, you can do extra formatting:

In [25]: df['Win_Rate'] = df['Win_Rate'].mul(100).apply(lambda x: '{}%'.format(int(x)))

In [26]: df
Out[26]:
   PlayerId  Status Win_Rate
0         1       1     100%
1         1       0      50%
2         1       0      33%
3         1       1      50%
4         2       1     100%
5         2       0      50%
6         2       0      33%
7         2       1      50%

Upvotes: 5

Related Questions