You are reading an old version of the documentation (v3.1.3). For the latest version see
Version 3.1.3
Fork me on GitHub

Table of Contents

Related Topics

Percentiles as horizontal bar chart

Bar charts are useful for visualizing counts, or summary statistics with error bars. Also see the Grouped bar chart with labels or the Horizontal bar chart example for simpler versions of those features.

This example comes from an application in which grade school gym teachers wanted to be able to show parents how their child did across a handful of fitness tests, and importantly, relative to how other children did. To extract the plotting code for demo purposes, we'll just make up some data for little Johnny Doe.

import numpy as np
import matplotlib
import matplotlib.pyplot as plt
from matplotlib.ticker import MaxNLocator
from collections import namedtuple


Student = namedtuple('Student', ['name', 'grade', 'gender'])
Score = namedtuple('Score', ['score', 'percentile'])

testNames = ['Pacer Test', 'Flexed Arm\n Hang', 'Mile Run', 'Agility',
             'Push Ups']
testMeta = dict(zip(testNames, ['laps', 'sec', 'min:sec', 'sec', '']))

def attach_ordinal(num):
    """helper function to add ordinal string to integers

    1 -> 1st
    56 -> 56th
    suffixes = {str(i): v
                for i, v in enumerate(['th', 'st', 'nd', 'rd', 'th',
                                       'th', 'th', 'th', 'th', 'th'])}

    v = str(num)
    # special case early teens
    if v in {'11', '12', '13'}:
        return v + 'th'
    return v + suffixes[v[-1]]

def format_score(scr, test):
    Build up the score labels for the right Y-axis by first
    appending a carriage return to each string and then tacking on
    the appropriate meta information (i.e., 'laps' vs 'seconds'). We
    want the labels centered on the ticks, so if there is no meta
    info (like for pushups) then don't add the carriage return to
    the string
    md = testMeta[test]
    if md:
        return '{0}\n{1}'.format(scr, md)
        return scr

def format_ycursor(y):
    y = int(y)
    if y < 0 or y >= len(testNames):
        return ''
        return testNames[y]

def plot_student_results(student, scores, cohort_size):
    #  create the figure
    fig, ax1 = plt.subplots(figsize=(9, 7))
    fig.subplots_adjust(left=0.115, right=0.88)
    fig.canvas.set_window_title('Eldorado K-8 Fitness Chart')

    pos = np.arange(len(testNames))

    rects = ax1.barh(pos, [scores[k].percentile for k in testNames],


    ax1.set_xlim([0, 100])
    ax1.xaxis.grid(True, linestyle='--', which='major',
                   color='grey', alpha=.25)

    # Plot a solid vertical gridline to highlight the median position
    ax1.axvline(50, color='grey', alpha=0.25)

    # Set the right-hand Y-axis ticks and labels
    ax2 = ax1.twinx()

    scoreLabels = [format_score(scores[k].score, k) for k in testNames]

    # set the tick locations
    # make sure that the limits are set equally on both yaxis so the
    # ticks line up

    # set the tick labels

    ax2.set_ylabel('Test Scores')

    xlabel = ('Percentile Ranking Across {grade} Grade {gender}s\n'
              'Cohort Size: {cohort_size}')

    rect_labels = []
    # Lastly, write in the ranking inside each bar to aid in interpretation
    for rect in rects:
        # Rectangle widths are already integer-valued but are floating
        # type, so it helps to remove the trailing decimal point and 0 by
        # converting width to int type
        width = int(rect.get_width())

        rankStr = attach_ordinal(width)
        # The bars aren't wide enough to print the ranking inside
        if width < 40:
            # Shift the text to the right side of the right edge
            xloc = 5
            # Black against white background
            clr = 'black'
            align = 'left'
            # Shift the text to the left side of the right edge
            xloc = -5
            # White on magenta
            clr = 'white'
            align = 'right'

        # Center the text vertically in the bar
        yloc = rect.get_y() + rect.get_height() / 2
        label = ax1.annotate(rankStr, xy=(width, yloc), xytext=(xloc, 0),
                            textcoords="offset points",
                            ha=align, va='center',
                            color=clr, weight='bold', clip_on=True)

    # make the interactive mouse over give the bar title
    ax2.fmt_ydata = format_ycursor
    # return all of the artists created
    return {'fig': fig,
            'ax': ax1,
            'ax_right': ax2,
            'bars': rects,
            'perc_labels': rect_labels}

student = Student('Johnny Doe', 2, 'boy')
scores = dict(zip(testNames,
                  (Score(v, p) for v, p in
                   zip(['7', '48', '12:52', '17', '14'],
                       np.round(np.random.uniform(0, 1,
                                                  len(testNames)) * 100, 0)))))
cohort_size = 62  # The number of other 2nd grade boys

arts = plot_student_results(student, scores, cohort_size)


The use of the following functions, methods and classes is shown in this example:


<function _AxesBase.twinx at 0x7f18a5a81670>

Keywords: matplotlib code example, codex, python plot, pyplot Gallery generated by Sphinx-Gallery