Log Transformation Reduces Normality of Volume Distribution

[OliShah]

Hi,

in the data loading section the Volume column goes through a log transformation, as per:

data["log_volume"] = np.log(data.volume + 1e-8)

I looked at a qqplot, a histogram and performed a Shapiro-Walk test before and after the transformation and it seems the transformation led to less normality.

I'm a novice at all this but i would suspect (after some conversations with chatgpt) that the transformation would show greater normality unlike the image below. Can anybody set me straight on this one?

Thanks in advance!