Interactive bar plot using matplotlib with annotation change

Question

I'm trying to create a horizontal bar plot that can dynamically change with a slider. I've followed the recipe on the matplotib website and it works well for line data. The code:

def interactive_pyramid(ages_matrix, year_labels):
    fig, axs = plt.subplots()
    plt.subplots_adjust(bottom=0.25)
    l, = axs.plot(range(len(ages_matrix[0, :])), ages_matrix[0, :])
    axs.annotate(year_labels[0], xy=(0.85, 0.85), xycoords="axes fraction")
    axs.set_ylim([0,800])

    pprint (dir(axs))
    axcolor = 'lightgoldenrodyellow'
    axyear = plt.axes([0.25, 0.1, 0.5, 0.1], axisbg=axcolor)
    syear = Slider(axyear, 'Year', 0, ages_matrix.shape[0] - 1, 0)

    def update(val):
        year = syear.val
        # axs.barh(range(len(ages_matrix[0, :])), ages_matrix[val, :])
        l.set_ydata(ages_matrix[val, :])
        # axs.annotate(year_labels[year], xy=(0.85, 0.85), xycoords="axes fraction")
        axs.Annotation.remove()
        fig.canvas.draw()
    syear.on_changed(update)

ages_matrix is a 2d ndarray and year_labels is a 1d ndarray

two main questions:

the axs.barh() does not return an object with a set_ydata() method do I can't change the y data. if I just draw the data again on the axs object it doesn't erase the previous information, resulting in a clutter of charts.
the same happens with annotations - it does not erase the previous one.

Is there any way to efficiently erase the ax and draw it again? maybe some way to refresh the canvas?

You need to call remove on the instance of the `Annotation` object. `barh` returns a list of rects, for now you have to change each of their height independently. — tacaswell, Sep 22 '15 at 13:52
Thanks, it kinda helped. Thought it would be more elegant though. — DeanLa, Sep 22 '15 at 15:04

score 1 · Answer 1 · answered Sep 22 '15 at 15:09

That's what I came up with:

# Init Chart
fig, axs = plt.subplots()
plt.subplots_adjust(bottom=0.25)
axs.set_xlim([0,ages_matrix.max()*1.05])
# Initial Pyramid
pyramid = axs.barh(np.arange(len(ages_matrix[0, :])) * 5,
                   ages_matrix[0, :],
                   height=4.5)
# Annotation
ann = axs.annotate(year_labels[0], xy=(0.85, 0.85), xycoords="axes fraction")
# Slider
axcolor = 'lightgoldenrodyellow'
axyear = plt.axes([0.25, 0.1, 0.5, 0.1], axisbg=axcolor)
syear = Slider(axyear, 'Year', 0, ages_matrix.shape[0] - 1, 0)

def update(val):
    t = syear.val
    year = np.trunc(year_labels[t])
    day = (year_labels[t] - year) * 365
    ages = ages_matrix[t, :]
    for i, p in enumerate(pyramid):
        p.set_width(ages[i])
    ann.set_text("Year: {}\nDay: {}".format(int(year), int(day)))
    fig.canvas.draw()

Interactive bar plot using matplotlib with annotation change

1 Answers1