Using hvplot pandas dataframe to create a histogram but passing parameter bins='auto' or bins='fd'

hvPlot
1

Hello community,

Trying to utilize the hvplot API to crate a panel dashboard and want to create a histogram but relying on numpy.histogram_bin_edges; auto or fd but getting an error when changing the backend from matplotlib to holoviews.

Hope you can point me in the right direction.

Here the toy example:

x = np.linspace(norm.ppf(0.01),
            norm.ppf(0.99), 100)
df_toy = pd.DataFrame(x,columns=['x'])
df_toy.describe()
df_toy['y'] = norm.pdf(df_toy['x'],df_toy['x'].mean(),df_toy['x'].std())

df_toy['x'].plot(kind='hist',bins='fd')

The error I got is:
TypeError: bins must be an integer, a string, or an array

If I utilize matplotlib works fine if using ‘fd’. With auto I got:

ValueError: shape mismatch: objects cannot be broadcast to a single shape

but if I create the chart using:

plt.hist(df_toy['x'],bins='auto')

That works fine with auto.

2

Which version of HoloViews are you using? The bins='auto'/'fd' options weren’t added until the very recent 1.13.3 release.

3

There you go! was using: holoviews 1.13.2.
Thank you!