I have a pandas dataframe and I'd like to calculate the percentage difference compared to the beginning value. Is there a way to use pct_change() to complete this task?
e.g
my data
date close
2020-05-08 100
2020-05-11 102
2020-05-12 108
I'd like the change to appear like
date close change (%)
2020-05-08 100 0
2020-05-11 102 2
2020-05-12 108 8
You can do
df['change'] = df['close'].div(df['close'].iloc[0]).sub(1).mul(100)
or
df['change'] = (df.close/df.close.iloc[0] - 1) * 100
Output:
date close change
0 2020-05-08 100 0.0
1 2020-05-11 102 2.0
2 2020-05-12 108 8.0
Yes, this can be done using pct_change().
df['change'] = (df['close'].pct_change().fillna(0).add(1).cumprod().sub(1))*100
If you don't care about direction (+/-), then you can also take the absolute value:
df['change'] = (df['close'].pct_change().fillna(0).add(1).cumprod().sub(1).abs())*100
Edit: as David suggested, there are better solutions than this. I will leave the answer so you guys and girls know what not to do :).
You can always use an extra column to help. It's fairly easy:
start_value = 100 #or access through the index
df['beginning'] = start_value
df['change (%)'] = 100 * (df.close/df.beginning - 1)
df
Output:
date close beginning change (%)
0 2020-05-08 100 100 0.0
1 2020-05-11 102 100 2.0
2 2020-05-12 108 100 8.0