Matplotlib subplots not working with seaborn distplot

[tommylees112]

Code sample, a copy-pastable example if possible

A "Minimal, Complete and Verifiable Example" will make it much easier for maintainers to help you.

import pandas as pd
import xarray as xr
import numpy as np


def drop_nans_and_flatten(dataArray: xr.DataArray) -> np.ndarray:
    """flatten the array and drop nans from that array. Useful for plotting histograms.

    Arguments:
    ---------
    : dataArray (xr.DataArray)
        the DataArray of your value you want to flatten
    """
    # drop NaNs and flatten
    return dataArray.values[~np.isnan(dataArray.values)]


# create dimensions of xarray object
times = pd.date_range(start='1981-01-31', end='2019-04-30', freq='M')
lat = np.linspace(0, 1, 224)
lon = np.linspace(0, 1, 176)

rand_arr = np.random.randn(len(times), len(lat), len(lon))

# create xr.Dataset
coords = {'time': times, 'lat':lat, 'lon':lon}
dims = ['time', 'lat', 'lon']
ds = xr.Dataset({'precip': (dims, rand_arr)}, coords=coords)
ds['month'], ds['year'] = ds['time.month'], ds['time.year']

Making plot using proplot

import proplot as plot
import calendar

f, axs = plot.subplots(nrows=4, ncols=3, axwidth=1.5, figsize=(8,12), share=2) # share=3, span=1,
axs.format(
    xlabel='Precip', ylabel='Density', suptitle='Distribution', 
)

month_abbrs = list(calendar.month_abbr)
mean_ds = ds.groupby('time.month').mean(dim='time')
flattened = []
for mth in np.arange(1, 13):
    ax = axs[mth - 1]
    ax.set_title(month_abbrs[mth])
    print(f"Plotting {month_abbrs[mth]}")
    flat = drop_nans_and_flatten(mean_ds.sel(month=mth).precip)
    flattened.append(flat)
    sns.distplot(flat, ax=ax, **{'kde': False})

Actual result vs. expected result

This should explain why the current behavior is a problem and why the expected result is a better solution.

The proplot returns a plot like follows:

It looks empty plot. Also the axes are only sharing the x-axis for each column but I want it to be shared across all subplots.

The matplotlib version does what I expect.

fig, axs = plt.subplots(4, 3, figsize=(8, 12), sharex=True, sharey=True)

month_abbrs = [m for m in calendar.month_abbr if m != '']

for mth in range(0, 12):
    ax_ix = np.unravel_index(mth, (4, 3))
    ax = axs[ax_ix]
    mth_str = month_abbrs[mth]
    sns.distplot(flattened[mth], ax=ax)
    ax.set_title(mth_str)

fig.suptitle('Distribution of Rainfall each Month');

[lukelbd]

Thanks for the examples (and your interest)! Never thought of using seaborn with ProPlot. This is probably related to my fill_between_wrapper overrides. All of ProPlot's plotting wrappers are supposed to be strict supersets/not interfer with usage of the existing matplotlib API, but I appear to have overlooked something. Will get to it within the next few days.

[tommylees112]

Oooooh interesting! I love the package idea because I myself have never got round to understanding the many ways of controlling matplotlib settings. Should probably look into that too! But yes if it turns out to be a pain ill use a different histogram function I just quite like the seaborn one and it’s easy to use! Also because it’s a big part of python visualisation I’m guessing others might want it too! Thanks so much

…

Sent from my iPhone

On 18 Jul 2019, at 23:22, Luke Davis ***@***.***> wrote: Thanks for the examples (and your interest)! Never thought of using seaborn with ProPlot. This is probably related to my fill_between_wrapper overrides. All of ProPlot's plotting wrappers are supposed to be strict supersets/not interfer with usage of the existing matplotlib API, but I appear to have overlooked something. Will get to it within the next few days. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[lukelbd]

Sorry for the delay -- the latest commit (c45c88a) fixes this. The source of the bug is complicated, but basically I was wrapping the matplotlib.axes.Axes.hist method improperly. Try running pip uninstall proplot then re-installing with pip install git+https://github.com/lukelbd/proplot.git#egg=proplot (when ProPlot is officially released it will use a versioning system, but for now I'm just pushing bugfixes directly to master).

Also note that you cannot specify both the figure dimensions and the axes dimensions. If you input the former the latter will be ignored. ProPlot now issues a warning when this happens.

[lukelbd]

And thanks again for posting the example, you made it a very quick fix.

[Ntkrell]

@tommylees112 maybe you can help me with this. When I try and run your code above I get ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all(). Did you ever run into this error?

I just installed proplot so I have version 0.5.0.

Here's the code I used:

import pandas as pd
import xarray as xr
import numpy as np
import seaborn as sns

import proplot as plot
import calendar

def drop_nans_and_flatten(dataArray: xr.DataArray) -> np.ndarray:
    """flatten the array and drop nans from that array. Useful for plotting histograms.

    Arguments:
    ---------
    : dataArray (xr.DataArray)
        the DataArray of your value you want to flatten
    """
    # drop NaNs and flatten
    return dataArray.values[~np.isnan(dataArray.values)]


# create dimensions of xarray object
times = pd.date_range(start='1981-01-31', end='2019-04-30', freq='M')
lat = np.linspace(0, 1, 224)
lon = np.linspace(0, 1, 176)

rand_arr = np.random.randn(len(times), len(lat), len(lon))

# create xr.Dataset
coords = {'time': times, 'lat':lat, 'lon':lon}
dims = ['time', 'lat', 'lon']
ds = xr.Dataset({'precip': (dims, rand_arr)}, coords=coords)
ds['month'], ds['year'] = ds['time.month'], ds['time.year']

f, axs = plot.subplots(nrows=4, ncols=3, axwidth=1.5, figsize=(8,12), share=2) # share=3, span=1,
axs.format(
    xlabel='Precip', ylabel='Density', suptitle='Distribution', 
)

month_abbrs = list(calendar.month_abbr)
mean_ds = ds.groupby('time.month').mean(dim='time')
flattened = []
for mth in np.arange(1, 13):
    ax = axs[mth - 1]
    ax.set_title(month_abbrs[mth])
    print(f"Plotting {month_abbrs[mth]}")
    flat = drop_nans_and_flatten(mean_ds.sel(month=mth).precip)
    flattened.append(flat)
    sns.distplot(flat, ax=ax, **{'kde': False})

And the error:

Plotting Jan
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-1-bea6a445aa3a> in <module>
     46     flat = drop_nans_and_flatten(mean_ds.sel(month=mth).precip)
     47     flattened.append(flat)
---> 48     sns.distplot(flat, ax=ax, **{'kde': False})
     49

~/miniconda3/envs/maize-Toff/lib/python3.7/site-packages/seaborn/distributions.py in distplot(a, bins, hist, kde, rug, fit, hist_kws, kde_kws, rug_kws, fit_kws, color, vertical, norm_hist, axlabel, label, ax)
    226         ax.hist(a, bins, orientation=orientation,
    227                 color=hist_color, **hist_kws)
--> 228         if hist_color != color:
    229             hist_kws["color"] = hist_color
    230

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Bringing this up because I'm trying to make a similar plot (proplot subplots + sns.distplot()) but am running into that ValueError for my code as well. Hoping it's an easy fix.