seaborn.histplot — seaborn 0.12.2 documentation (2023)

seaborn.histplot(data=None, *, x=None, y=None, hue=None, weights=None, stat='count', bins='auto', binwidth=None, binrange=None, discrete=None, cumulative=False, common_bins=True, common_norm=True, multiple='layer', element='bars', fill=True, shrink=1, kde=False, kde_kws=None, line_kws=None, thresh=0, pthresh=None, pmax=None, cbar=False, cbar_ax=None, cbar_kws=None, palette=None, hue_order=None, hue_norm=None, color=None, log_scale=None, legend=True, ax=None, **kwargs)#

Plot univariate or bivariate histograms to show distributions of datasets.

A histogram is a classic visualization tool that represents the distributionof one or more variables by counting the number of observations that fall withindiscrete bins.

This function can normalize the statistic computed within each bin to estimatefrequency, density or probability mass, and it can add a smooth curve obtainedusing a kernel density estimate, similar to kdeplot().

More information is provided in the user guide.

datapandas.DataFrame, numpy.ndarray, mapping, or sequence

Input data structure. Either a long-form collection of vectors that can beassigned to named variables or a wide-form dataset that will be internallyreshaped.

x, yvectors or keys in data

Variables that specify positions on the x and y axes.

huevector or key in data

Semantic variable that is mapped to determine the color of plot elements.

weightsvector or key in data

If provided, weight the contribution of the corresponding data pointstowards the count in each bin by these factors.


Aggregate statistic to compute in each bin.

  • count: show the number of observations in each bin

  • frequency: show the number of observations divided by the bin width

  • probability or proportion: normalize such that bar heights sum to 1

  • percent: normalize such that bar heights sum to 100

  • density: normalize such that the total area of the histogram equals 1

binsstr, number, vector, or a pair of such values

Generic bin parameter that can be the name of a reference rule,the number of bins, or the breaks of the bins.Passed to numpy.histogram_bin_edges().

binwidthnumber or pair of numbers

Width of each bin, overrides bins but can be used withbinrange.

binrangepair of numbers or a pair of pairs

Lowest and highest value for bin edges; can be used eitherwith bins or binwidth. Defaults to data extremes.


If True, default to binwidth=1 and draw the bars so that they arecentered on their corresponding data points. This avoids “gaps” that mayotherwise appear when using discrete (integer) data.


If True, plot the cumulative counts as bins increase.


If True, use the same bins when semantic variables produce multipleplots. If using a reference rule to determine the bins, it will be computedwith the full dataset.


If True and using a normalized statistic, the normalization will apply overthe full dataset. Otherwise, normalize each histogram independently.

multiple{“layer”, “dodge”, “stack”, “fill”}

Approach to resolving multiple elements when semantic mapping creates subsets.Only relevant with univariate data.

element{“bars”, “step”, “poly”}

Visual representation of the histogram statistic.Only relevant with univariate data.


If True, fill in the space under the histogram.Only relevant with univariate data.


Scale the width of each bar relative to the binwidth by this factor.Only relevant with univariate data.


If True, compute a kernel density estimate to smooth the distributionand show on the plot as (one or more) line(s).Only relevant with univariate data.


Parameters that control the KDE computation, as in kdeplot().


Parameters that control the KDE visualization, passed tomatplotlib.axes.Axes.plot().

threshnumber or None

Cells with a statistic less than or equal to this value will be transparent.Only relevant with bivariate data.

pthreshnumber or None

Like thresh, but a value in [0, 1] such that cells with aggregate counts(or other statistics, when used) up to this proportion of the total will betransparent.

pmaxnumber or None

A value in [0, 1] that sets that saturation point for the colormap at a valuesuch that cells below constitute this proportion of the total count (orother statistic, when used).


If True, add a colorbar to annotate the color mapping in a bivariate plot.Note: Does not currently support plots with a hue variable well.


Pre-existing axes for the colorbar.


Additional parameters passed to matplotlib.figure.Figure.colorbar().

palettestring, list, dict, or matplotlib.colors.Colormap

Method for choosing the colors to use when mapping the hue semantic.String values are passed to color_palette(). List or dict valuesimply categorical mapping, while a colormap object implies numeric mapping.

hue_ordervector of strings

Specify the order of processing and plotting for categorical levels of thehue semantic.

hue_normtuple or matplotlib.colors.Normalize

Either a pair of values that set the normalization range in data unitsor an object that will map from data units into a [0, 1] interval. Usageimplies numeric mapping.

colormatplotlib color

Single color specification for when hue mapping is not used. Otherwise, theplot will try to hook into the matplotlib property cycle.

log_scalebool or number, or pair of bools or numbers

Set axis scale(s) to log. A single value sets the data axis for univariatedistributions and both axes for bivariate distributions. A pair of valuessets each axis independently. Numeric values are interpreted as the desiredbase (default 10). If False, defer to the existing Axes scale.


If False, suppress the legend for semantic variables.


Pre-existing axes for the plot. Otherwise, call matplotlib.pyplot.gca()internally.


Other keyword arguments are passed to one of the following matplotlibfunctions:


The matplotlib axes containing the plot.

See also


Figure-level interface to distribution plot functions.


Plot univariate or bivariate distributions using kernel density estimation.


Plot a tick at each observation value along the x and/or y axes.


Plot empirical cumulative distribution functions.


Draw a bivariate plot with univariate marginal distributions.


The choice of bins for computing and plotting a histogram can exertsubstantial influence on the insights that one is able to draw from thevisualization. If the bins are too large, they may erase important features.On the other hand, bins that are too small may be dominated by randomvariability, obscuring the shape of the true underlying distribution. Thedefault bin size is determined using a reference rule that depends on thesample size and variance. This works well in many cases, (i.e., with“well-behaved” data) but it fails in others. It is always a good to trydifferent bin sizes to be sure that you are not missing something important.This function allows you to specify bins in several different ways, such asby setting the total number of bins to use, the width of each bin, or thespecific locations where the bins should break.


Assign a variable to x to plot a univariate distribution along the x axis:

penguins = sns.load_dataset("penguins")sns.histplot(data=penguins, x="flipper_length_mm")
seaborn.histplot — seaborn 0.12.2 documentation (1)

Flip the plot by assigning the data variable to the y axis:

sns.histplot(data=penguins, y="flipper_length_mm")
seaborn.histplot — seaborn 0.12.2 documentation (2)

Check how well the histogram represents the data by specifying adifferent bin width:

sns.histplot(data=penguins, x="flipper_length_mm", binwidth=3)
seaborn.histplot — seaborn 0.12.2 documentation (3)

You can also define the total number of bins to use:

sns.histplot(data=penguins, x="flipper_length_mm", bins=30)
seaborn.histplot — seaborn 0.12.2 documentation (4)

Add a kernel density estimate to smooth the histogram, providingcomplementary information about the shape of the distribution:

sns.histplot(data=penguins, x="flipper_length_mm", kde=True)
seaborn.histplot — seaborn 0.12.2 documentation (5)

If neither x nor y is assigned, the dataset is treated aswide-form, and a histogram is drawn for each numeric column:

seaborn.histplot — seaborn 0.12.2 documentation (6)

You can otherwise draw multiple histograms from a long-form dataset withhue mapping:

sns.histplot(data=penguins, x="flipper_length_mm", hue="species")
seaborn.histplot — seaborn 0.12.2 documentation (7)

The default approach to plotting multiple distributions is to “layer”them, but you can also “stack” them:

sns.histplot(data=penguins, x="flipper_length_mm", hue="species", multiple="stack")
seaborn.histplot — seaborn 0.12.2 documentation (8)

Overlapping bars can be hard to visually resolve. A different approachwould be to draw a step function:

sns.histplot(penguins, x="flipper_length_mm", hue="species", element="step")
seaborn.histplot — seaborn 0.12.2 documentation (9)

You can move even farther away from bars by drawing a polygon withvertices in the center of each bin. This may make it easier to see theshape of the distribution, but use with caution: it will be less obviousto your audience that they are looking at a histogram:

sns.histplot(penguins, x="flipper_length_mm", hue="species", element="poly")
seaborn.histplot — seaborn 0.12.2 documentation (10)

To compare the distribution of subsets that differ substantially insize, use independent density normalization:

sns.histplot( penguins, x="bill_length_mm", hue="island", element="step", stat="density", common_norm=False,)
seaborn.histplot — seaborn 0.12.2 documentation (11)

It’s also possible to normalize so that each bar’s height shows aprobability, proportion, or percent, which make more sense for discretevariables:

tips = sns.load_dataset("tips")sns.histplot(data=tips, x="size", stat="percent", discrete=True)
seaborn.histplot — seaborn 0.12.2 documentation (12)

You can even draw a histogram over categorical variables (although thisis an experimental feature):

sns.histplot(data=tips, x="day", shrink=.8)
seaborn.histplot — seaborn 0.12.2 documentation (13)

When using a hue semantic with discrete data, it can make sense to“dodge” the levels:

sns.histplot(data=tips, x="day", hue="sex", multiple="dodge", shrink=.8)
seaborn.histplot — seaborn 0.12.2 documentation (14)

Real-world data is often skewed. For heavily skewed distributions, it’s better to define the bins in log space. Compare:

planets = sns.load_dataset("planets")sns.histplot(data=planets, x="distance")
seaborn.histplot — seaborn 0.12.2 documentation (15)

To the log-scale version:

sns.histplot(data=planets, x="distance", log_scale=True)
seaborn.histplot — seaborn 0.12.2 documentation (16)

There are also a number of options for how the histogram appears. Youcan show unfilled bars:

sns.histplot(data=planets, x="distance", log_scale=True, fill=False)
seaborn.histplot — seaborn 0.12.2 documentation (17)

Or an unfilled step function:

sns.histplot(data=planets, x="distance", log_scale=True, element="step", fill=False)
seaborn.histplot — seaborn 0.12.2 documentation (18)

Step functions, esepcially when unfilled, make it easy to comparecumulative histograms:

sns.histplot( data=planets, x="distance", hue="method", hue_order=["Radial Velocity", "Transit"], log_scale=True, element="step", fill=False, cumulative=True, stat="density", common_norm=False,)
seaborn.histplot — seaborn 0.12.2 documentation (19)

When both x and y are assigned, a bivariate histogram iscomputed and shown as a heatmap:

sns.histplot(penguins, x="bill_depth_mm", y="body_mass_g")
seaborn.histplot — seaborn 0.12.2 documentation (20)

It’s possible to assign a hue variable too, although this will notwork well if data from the different levels have substantial overlap:

sns.histplot(penguins, x="bill_depth_mm", y="body_mass_g", hue="species")
seaborn.histplot — seaborn 0.12.2 documentation (21)

Multiple color maps can make sense when one of the variables isdiscrete:

sns.histplot( penguins, x="bill_depth_mm", y="species", hue="species", legend=False)
seaborn.histplot — seaborn 0.12.2 documentation (22)

The bivariate histogram accepts all of the same options for computationas its univariate counterpart, using tuples to parametrize x andy independently:

sns.histplot( planets, x="year", y="distance", bins=30, discrete=(True, False), log_scale=(False, True),)
seaborn.histplot — seaborn 0.12.2 documentation (23)

The default behavior makes cells with no observations transparent,although this can be disabled:

sns.histplot( planets, x="year", y="distance", bins=30, discrete=(True, False), log_scale=(False, True), thresh=None,)
seaborn.histplot — seaborn 0.12.2 documentation (24)

It’s also possible to set the threshold and colormap saturation point interms of the proportion of cumulative counts:

sns.histplot( planets, x="year", y="distance", bins=30, discrete=(True, False), log_scale=(False, True), pthresh=.05, pmax=.9,)
seaborn.histplot — seaborn 0.12.2 documentation (25)

To annotate the colormap, add a colorbar:

sns.histplot( planets, x="year", y="distance", bins=30, discrete=(True, False), log_scale=(False, True), cbar=True, cbar_kws=dict(shrink=.75),)
seaborn.histplot — seaborn 0.12.2 documentation (26)


Top Articles
Latest Posts
Article information

Author: Chrissy Homenick

Last Updated: 11/13/2023

Views: 6043

Rating: 4.3 / 5 (54 voted)

Reviews: 85% of readers found this page helpful

Author information

Name: Chrissy Homenick

Birthday: 2001-10-22

Address: 611 Kuhn Oval, Feltonbury, NY 02783-3818

Phone: +96619177651654

Job: Mining Representative

Hobby: amateur radio, Sculling, Knife making, Gardening, Watching movies, Gunsmithing, Video gaming

Introduction: My name is Chrissy Homenick, I am a tender, funny, determined, tender, glorious, fancy, enthusiastic person who loves writing and wants to share my knowledge and understanding with you.