Errorbars

Download this notebook from GitHub (right-click to download).

---

import hvplot.pandas  # noqa

errorbars are most helpful when overlaid with other plot types. To do this we’ll use the * operator.

from bokeh.sampledata import perceptions

data = perceptions.numberly.describe().transpose().sort_values('mean')
data

	count	mean	std	min	25%	50%	75%	max
A couple	46.0	2.326087	1.033964	2.00	2.0000	2.000	2.000	8.0
A few	46.0	3.608696	1.357747	2.00	3.0000	3.000	4.000	10.0
Some	46.0	4.739130	2.123255	2.00	3.2500	4.500	5.000	15.0
Fractions of	46.0	5.109783	16.847218	0.01	0.1625	0.365	0.875	100.0
Several	46.0	6.304348	2.096235	2.00	5.0000	7.000	7.750	10.0
Dozens	46.0	35.130435	17.119980	12.00	24.0000	36.000	36.000	120.0
Many	46.0	44.065217	145.464223	5.00	9.2500	20.000	25.000	1000.0
A lot	46.0	85.065217	431.917451	7.00	10.0000	14.500	22.250	2948.0
Hundreds of	46.0	400.804348	424.387015	100.00	212.5000	300.000	475.000	3000.0
Scores of	46.0	3017.456522	15266.869723	3.00	40.0000	80.000	200.000	100000.0

First we’ll use just one column to set the size of the errorbars around ‘mean’.

data.hvplot.scatter(y='mean', ylabel='amount') * data.hvplot.errorbars(y='mean', yerr1='std')

Then we’ll use two columns. Remember that these are not the absolute placements of the ends of the bars, but the distance from the center.

data['yerr1'] = data['mean'] - data['min']
data['yerr2'] = data['max'] - data['mean']
data.hvplot.scatter(y='mean', ylabel='amount') * data.hvplot.errorbars(y='mean', yerr1='yerr1', yerr2='yerr2')

This web page was generated from a Jupyter notebook and not all interactivity will work on this website. Right click to download and run locally for full Python-backed interactivity.

---

Download this notebook from GitHub (right-click to download).