Donbeo - 5 months ago 72x

Python Question

I have a dataframe and I would like to plot it as:

`>>> X = pd.DataFrame(np.random.normal(0, 1, (100, 3)))`

>>> X['NCP'] = np.random.randint(0, 5, 100)

>>> X[X['NCP'] == 0] += 100

>>> X.groupby('NCP').boxplot()

The result is what I want but all the subplots have the same ylim. This makes impossible to visualize the result properly. How can I set different ylim for each subplot?

Answer

What you asked for was to set the y axis separately for each axes. I believe that should be `ax.set_ylim([a, b])`

. But every time I ran it for each axes it updated for all.

Because I couldn't figure out how to answer your question directly, I'm providing a work around.

```
X = pd.DataFrame(np.random.normal(0, 1, (100, 3)))
X['NCP'] = np.random.randint(0, 5, 100)
X[X['NCP'] == 0] += 100
groups = X.groupby('NCP')
print groups.groups.keys()
# This gets a number of subplots equal to the number of groups in a single
# column. you can adjust this yourself if you need.
fig, axes = plt.subplots(len(groups.groups), 1, figsize=[10, 12])
# Loop through each group and plot boxplot to appropriate axis
for i, k in enumerate(groups.groups.keys()):
group = groups.get_group(k)
group.boxplot(ax=axes[i], return_type='axes')
```

Source (Stackoverflow)

Comments