Matplotlib stick plot recipes

stickplot1.py

"""
Credits
-------
Stephane Raynaud 
http://permalink.gmane.org/gmane.comp.python.matplotlib.general/24155
"""
import matplotlib.pyplot as plt
import numpy as np
import datetime as dtime
from matplotlib.dates import date2num

""" fake dates starting now """
x = np.arange(100, 110, 0.1)
start = dtime.datetime.now()
dates = [start + dtime.timedelta(days=n) for n in range(len(x))]

""" dummy u, v """
u = np.sin(x)
v = np.cos(x)

fig, ax = plt.subplots(1, 1, figsize=(16,6))

qiv = ax.quiver(date2num(dates), [[0]*len(x)], u, v, headlength=0, headwidth=0, headaxislength=0 )
key = ax.quiverkey(qiv, 0.25, 0.75, 0.5, "0.5 N m$^{-2}$", labelpos='N', coordinates='axes' )
plt.setp(ax.get_yticklabels(), visible=False)
ax.xaxis_date()

plt.show()

stickplot2.py

"""
some code to do a stick plot with MPL

This version uses matplotlib's quiver.

Credits
-------
Chris Barker
http://matplotlib.1069221.n5.nabble.com/Stick-Plot-td21479.html

See also
--------
http://matplotlib.1069221.n5.nabble.com/scaling-arrows-in-quiver-td23758.html
"""

import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np

import datetime


def stick(axes, times, speeds, directions, units=''):
    """
    Create a stick plot of the given data on the given axes.
   
    Stick plots are commonly used to display time series of
    a vector quantity at a point, such as wind or ocean current observations.
    
    Call signature::
        stick(axes, times, speed, direction, units='m/s')
   
    Arguments: 
       *axes*:  The axes object you want to plot on
       *times*: An array of datetime objects giving the time of the
                observations
       *speeds*: An array of the velocities of the observations
       *directions*: An array of the directions of the observations (in
                     degrees from North, where North is up on the plot)
       *units*:  A string with the units of the observation
    
    """

    props = {'units' : "dots",
             'width' : 2,
             'headwidth': 0,
             'headlength': 0,
             'headaxislength': 0,
             'scale' : .1,
             }

    ##fixme: this should use some smarts to fit the data...
    label_scale = 10
    unit_label = "%3g %s"%(label_scale, units)

    y = np.zeros_like(speeds)
    dir_rad = directions / 180. * np.pi
    u = np.sin(dir_rad) * speeds
    v = np.cos(dir_rad) * speeds

    Q = ax.quiver(times, y, u, v, **props)
    ax.quiverkey(Q, X=0.1, Y=0.95, U=label_scale, label=unit_label, coordinates='axes', labelpos='S')
    yaxis = ax.yaxis
    yaxis.set_ticklabels([])

if __name__ == '__main__':

    ## some sample data:

    data = np.array([(10,  0.5),
                     (10,  22.5),
                     (10, 45.0),
                     (10, 67.5),
                     (10, 90.0),
                     (10, 112.5),
                     (10, 135.0),
                     (10, 157.5),
                     (10, 180.0),
                     (10, 202.5),
                     (10, 225.0),
                     (10, 247.5),
                     (10, 270.0),
                     (10, 292.5),
                     (10, 315.0),
                     (10, 337.5),
                     ( 8, 157.5),
                     (12, 180.0),
                     (15, 202.5),
                     ( 7, 225.0),
                     ( 5, 247.5),
                     (16, 270.0),
                     (20, 292.5),
                     (22, 315.0),
                     (15, 337.5),
                     (12,  22.5),
                     (17, 25.0),
                     (18, 30),
                     ], dtype=np.float)
    data = np.r_[data, data, data, data, data, data]
    speeds  = data[:,0]
    directions  = data[:,1]
    times = range(len(speeds))
    #times = [datetime.datetime(2009, 5, 13, 0) + datetime.timedelta(hours=h) for h in times]

    fig = plt.figure(1)
    fig.clear()

    ax = fig.add_subplot(1,1,1)
    stick(ax, times, speeds, directions, units='knots')

    plt.draw()
    plt.show()