I'd like to color my histogram according to my palette. Here's the code I used to make this, and here's the error I received when I tried an answer I found on here.
g = sns.jointplot(data=emb_df, x='f0', y='y', kind="hist", hue='klabels', palette='tab10', marginal_kws={'hist_kws': {'palette': 'tab10'}})
plt.show()
UserWarning: The marginal plotting function has changed to `histplot`, which does not accept the following argument(s): hist_kws.
I have also tried this:
plt.setp(g.ax_marg_y.patches, color='grey')
But this does not color my histogram according my 'klabels' parameter, just a flat grey.
The marginal plot is colored by default using the same palette with corresponding hue. So, you could just run it without marginal_kws=. The marginal_kws= go directly to the histplot; instead of marginal_kws={'hist_kws': {'palette': 'tab10'}}, the correct use would be marginal_kws={'palette': 'tab10'}. If you would like stacked bars, you could try marginal_kws={'multiple': 'stack'})
If you want the marginal plots to be larger, the ratio= parameter can be altered. The default is 5, meaning the central plot is 5 times as large as the marginal plots.
Here is an example:
from matplotlib import pyplot as plt
import seaborn as sns
iris = sns.load_dataset('iris')
g = sns.jointplot(data=iris, x='petal_length', y='sepal_length', kind="hist", hue='species', palette='tab10',
ratio=2, marginal_kws={'multiple': 'stack'})
sns.move_legend(g.ax_joint, loc='upper left') # optionally move the legend; seaborn >= 0.11.2 needed
plt.show()
To have these plots side-by-side as subplots, you can call the underlying sns.histplot either with both x= and y= filled in (2D histogram), only x= given (horizontal histogram) or only y= given (vertical histogram).
from matplotlib import pyplot as plt
import seaborn as sns
iris = sns.load_dataset('iris')
fig, (ax1, ax2, ax3) = plt.subplots(ncols=3, figsize=(15, 4))
sns.histplot(data=iris, x='petal_length', y='sepal_length', hue='species', palette='tab10', legend=False, ax=ax1)
sns.histplot(data=iris, x='petal_length', hue='species', palette='tab10', multiple='stack', legend=False, ax=ax2)
sns.histplot(data=iris, y='sepal_length', hue='species', palette='tab10', multiple='stack', ax=ax3)
sns.move_legend(ax3, bbox_to_anchor=[1.01, 1.01], loc='upper left')
plt.tight_layout()
plt.show()
Related
Running this below code produces seaborn facetgrid graphs.
merged1=merged[merged['TEST'].isin(['VL'])]
merged2=merged[merged['TEST'].isin(['CD4'])]
g = sns.relplot(data=merged1, x='Days Post-ART', y='Log of VL and CD4', col='PATIENT ID',col_wrap=4, kind="line", height=4, aspect=1.5,
color='b', facet_kws={'sharey':True,'sharex':True})
for patid, ax in g.axes_dict.items(): # axes_dict is new in seaborn 0.11.2
ax1 = ax.twinx()
sns.lineplot(data=merged2[merged2['PATIENT ID'] == patid], x='Days Post-ART', y='Log of VL and CD4', color='r')
I've used the facet_kws={'sharey':True, 'sharex':True} to share the x-axis and y-axis but it's not working properly. Can someone assist?
As stated in the comments, the FacetGrid axes are shared by default. However, the twinx axes are not. Also, the call to twinx seems to reset the default hiding of the y tick labels.
You can manually share the twinx axes, and remove the unwanted tick labels.
Here is some example code using the iris dataset:
from matplotlib import pyplot as plt
import seaborn as sns
import numpy as np
iris = sns.load_dataset('iris')
g = sns.relplot(data=iris, x='petal_length', y='petal_width', col='species', col_wrap=2, kind="line",
height=4, aspect=1.5, color='b')
last_axes = np.append(g.axes.flat[g._col_wrap - 1::g._col_wrap], g.axes.flat[-1])
shared_right_y = None
for species, ax in g.axes_dict.items():
ax1 = ax.twinx()
if shared_right_y is None:
shared_right_y = ax1
else:
shared_right_y.get_shared_y_axes().join(shared_right_y, ax1)
sns.lineplot(data=iris[iris['species'] == species], x='petal_length', y='sepal_length', color='r', ax=ax1)
if not ax in last_axes: # remove tick labels from secondary axis
ax1.yaxis.set_tick_params(labelleft=False, labelright=False)
ax1.set_ylabel('')
if not ax in g._left_axes: # remove tick labels from primary axis
ax.yaxis.set_tick_params(labelleft=False, labelright=False)
plt.tight_layout()
plt.show()
In version 3.4, matplotlib added automatic Bar labels:
https://matplotlib.org/stable/users/whats_new.html#new-automatic-labeling-for-bar-charts
I'm trying to use this on a bar plot generated by Seaborn.
fig, axs = plt.subplots(
nrows=2,
)
for i, col in enumerate(['col_1', 'col_2']):
ax = axs[i]
sns.barplot(
x="class",
y=col,
hue="hue_col",
data=data_df,
edgecolor=".3",
linewidth=0.5,
ax=ax
)
ax.bar_label(ax.containers[i]) # Doesn't work
What do I need to do to make this work? example plot
You can loop through the containers and call ax.bar_label(...) for each of them. Note that seaborn creates one set of bars for each hue value.
The following example uses the titanic dataset and sets ci=None to avoid the error bars overlapping with the text (if error bars are needed, one could set a lighter color, e.g. errcolor='gold').
import seaborn as sns
import matplotlib.pyplot as plt
titanic = sns.load_dataset('titanic')
fig, axs = plt.subplots(ncols=2, figsize=(12, 4))
for ax, col in zip(axs, ['age', 'fare']):
sns.barplot(
x='sex',
y=col,
hue="class",
data=titanic,
edgecolor=".3",
linewidth=0.5,
ci=None,
ax=ax
)
ax.set_title('mean ' + col)
ax.margins(y=0.1) # make room for the labels
for bars in ax.containers:
ax.bar_label(bars, fmt='%.1f')
plt.tight_layout()
plt.show()
I would like to make a paired histogram like the one shown here using the seaborn distplot.
This kind of plot can also be referred to as the back-to-back histogram shown here, or a bihistogram inverted/mirrored along the x-axis as discussed here.
Here is my code:
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
green = np.random.normal(20,10,1000)
blue = np.random.poisson(60,1000)
fig, ax = plt.subplots(figsize=(8,6))
sns.distplot(blue, hist=True, kde=True, hist_kws={'edgecolor':'black'}, kde_kws={'linewidth':2}, bins=10, color='blue')
sns.distplot(green, hist=True, kde=True, hist_kws={'edgecolor':'black'}, kde_kws={'linewidth':2}, bins=10, color='green')
ax.set_xticks(np.arange(-20,121,20))
ax.set_yticks(np.arange(0.0,0.07,0.01))
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.show()
Here is the output:
When I use the method discussed here (plt.barh), I get the bar plot shown just below, which is not what I am looking for.
Or maybe I haven't understood the workaround well enough...
A simple/short implementation of python-seaborn-distplot similar to these kinds of plots would be perfect. I edited the figure of my first plot above to show the kind of plot I hope to achieve (though y-axis not upside down):
Any leads would be greatly appreciated.
You could use two subplots and invert the y-axis of the lower one and plot with the same bins.
df = pd.DataFrame({'a': np.random.normal(0,5,1000), 'b': np.random.normal(20,5,1000)})
fig =plt.figure(figsize=(5,5))
ax = fig.add_subplot(211)
ax2 = fig.add_subplot(212)
bins = np.arange(-20,40)
ax.hist(df['a'], bins=bins)
ax2.hist(df['b'],color='orange', bins=bins)
ax2.invert_yaxis()
edit:
improvements suggested by #mwaskom
fig, axes = plt.subplots(nrows=2, ncols=1, sharex=True, figsize=(5,5))
bins = np.arange(-20,40)
for ax, column, color, invert in zip(axes.ravel(), df.columns, ['teal', 'orange'], [False,True]):
ax.hist(df[column], bins=bins, color=color)
if invert:
ax.invert_yaxis()
plt.subplots_adjust(hspace=0)
Here is a possible approach using seaborn's displots.
Seaborn doesn't return the created graphical elements, but the ax can be interrogated. To make sure the ax only contains the elements you want upside down, those elements can be drawn first. Then, all the patches (the rectangular bars) and the lines (the curve for the kde) can be given their height in negative. Optionally the x-axis can be set at y == 0 using ax.spines['bottom'].set_position('zero').
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
green = np.random.normal(20, 10, 1000)
blue = np.random.poisson(60, 1000)
fig, ax = plt.subplots(figsize=(8, 6))
sns.distplot(green, hist=True, kde=True, hist_kws={'edgecolor': 'black'}, kde_kws={'linewidth': 2}, bins=10,
color='green')
for p in ax.patches: # turn the histogram upside down
p.set_height(-p.get_height())
for l in ax.lines: # turn the kde curve upside down
l.set_ydata(-l.get_ydata())
sns.distplot(blue, hist=True, kde=True, hist_kws={'edgecolor': 'black'}, kde_kws={'linewidth': 2}, bins=10,
color='blue')
ax.set_xticks(np.arange(-20, 121, 20))
ax.set_yticks(np.arange(0.0, 0.07, 0.01))
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
pos_ticks = np.array([t for t in ax.get_yticks() if t > 0])
ticks = np.concatenate([-pos_ticks[::-1], [0], pos_ticks])
ax.set_yticks(ticks)
ax.set_yticklabels([f'{abs(t):.2f}' for t in ticks])
ax.spines['bottom'].set_position('zero')
plt.show()
I have created a simple violin plot from a bands DataFrame (df10 below) using seaborn:
fig, ax = plt.subplots(figsize=(10,4))
ax = sns.violinplot(x='z', y='z_fit', hue='new_col', data=df10, cut=0, palette='Blues', linewidth=1)
ax.set_xlabel('z_sim')
ax.legend()
The legend is plotted automatically with the values of the hue parameter. Using ax.legend() I can only hide the name of the used column ('new_col').
However, I was wondering if there is some way to manually modify the legend (texts, colors and shapes) plotted below:
Example:
import seaborn as sns
tips = sns.load_dataset("tips")
g = sns.FacetGrid(tips, col="time", size=4, aspect=.75)
g = g.map(sns.violinplot, "sex", "total_bill", "smoker", palette={"No": "b", "Yes": "w"}, inner=None, linewidth=1, scale="area", split=True, width=0.75).despine(left=True)
g.fig.get_axes()[0].legend(title= 'smoker',loc='top left',labels=["YES","NO"],edgecolor='red',facecolor='blue',ncol=2)
g.set_axis_labels('lunch','total bill')
For more info run:
help(g.fig.get_axes()[0].legend)
Want labels for Bollinger Bands (R) ('upper band', 'rolling mean', 'lower band') to show up in legend. But legend just applies the same label to each line with the pandas label for the first (only) column, 'IBM'.
# Plot price values, rolling mean and Bollinger Bands (R)
ax = prices['IBM'].plot(title="Bollinger Bands")
rm_sym.plot(label='Rolling mean', ax=ax)
upper_band.plot(label='upper band', c='r', ax=ax)
lower_band.plot(label='lower band', c='r', ax=ax)
#
# Add axis labels and legend
ax.set_xlabel("Date")
ax.set_ylabel("Adjusted Closing Price")
ax.legend(loc='upper left')
plt.show()
I know this code may represent a fundamental lack of understanding of how matlibplot works so explanations are particularly welcome.
The problem is most probably that whatever upper_band and lower_band are, they are not labeled.
One option is to label them by putting them as column to a dataframe. This will allow to plot the dataframe column directly.
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
y =np.random.rand(4)
yupper = y+0.2
ylower = y-0.2
df = pd.DataFrame({"price" : y, "upper": yupper, "lower": ylower})
fig, ax = plt.subplots()
df["price"].plot(label='Rolling mean', ax=ax)
df["upper"].plot(label='upper band', c='r', ax=ax)
df["lower"].plot(label='lower band', c='r', ax=ax)
ax.legend(loc='upper left')
plt.show()
Otherwise you can also plot the data directly.
import matplotlib.pyplot as plt
import numpy as np
y =np.random.rand(4)
yupper = y+0.2
ylower = y-0.2
fig, ax = plt.subplots()
ax.plot(y, label='Rolling mean')
ax.plot(yupper, label='upper band', c='r')
ax.plot(ylower, label='lower band', c='r')
ax.legend(loc='upper left')
plt.show()
In both cases, you'll get a legend with labels. If that isn't enough, I recommend reading the Matplotlib Legend Guide which also tells you how to manually add labels to legends.