Python: Introduction to Data Visualization

The programs used in this tutorial are contained in DataVisualization.zip.

The programs used to teach the basics of plotting with Python are included in the Basics section below.

Programs that are used to illustrate more advanced concepts are linked here:

 

Basics

  • 01.py 
    from matplotlib.pyplot import figure, show

def main():
    #This first program will create a figure with nothing in it.
    figure()
    show()

if __name__ == "__main__":
    main()
  • 02.py 
    from matplotlib.pyplot import figure, show, plot

def main():
    #Building on the previous program, plot a line
    figure()

    plot([4,5,6])

    show()


if __name__ == "__main__":
    main()
  • 03.py
    from matplotlib.pyplot import figure, show, plot

def main():
    #Building on the previous program, plot a non-straight line
    figure()

    xdata=[1,2,3,4,5]
    ydata=[2,4,8,16,32]

    plot(xdata,ydata)

    show()


if __name__ == "__main__":
    main()
  • 04.py
    from matplotlib.pyplot import figure, show, plot, grid

def main():
    #Building on the previous program, plot a non-straight line
    figure()

    xdata=[1,2,3,4,5]
    ydata=[2,4,8,16,32]

    plot(xdata,ydata)

    show()


if __name__ == "__main__":
    main()
  • 05.py
    from matplotlib.pyplot import figure, show, plot, grid

def main():
    #Building on the previous program, plot a non-straight line and add a grid to the figure
    figure()

    xdata=[1,2,3,4,5]
    ydata=[10,5,30,10,12]

    plot(xdata,ydata)
    grid()
    
    show()


if __name__ == "__main__":
    main()
  • 06.py
    from matplotlib.pyplot import figure, show, plot, grid

from numpy import linspace, sin

def main():
    #Building on the previous program, plot a sin wave
    figure()

    xdata=linspace(0,10,100)
    sindata=sin(xdata)

    plot(xdata,sindata)
    grid()

    show()


if __name__ == "__main__":
    main()
  • 07.py
    from matplotlib.pyplot import figure, show, plot, grid, xlabel, ylabel, title

from numpy import linspace, sin

def main():
    #Building on the previous program, add titles and labels
    figure()

    xdata=linspace(0,20,200)
    sindata=sin(xdata)

    plot(xdata,sindata)
    grid()

    xlabel("x")
    ylabel("sin(x)")

    title("sin(x) plotted for many values of x")

    show()


if __name__ == "__main__":
    main()
  • 08.py
    from matplotlib.pyplot import figure, show, plot, grid, xlabel, ylabel, title

from numpy import linspace, sin, cos

def main():
    #Building on the previous program, add a second plot and change the labels and title
    figure()

    xdata=linspace(0,10,100)
    sindata=sin(xdata)
    cosdata=cos(xdata)

    plot(xdata,sindata)
    grid()

    xlabel("x")
    ylabel("sin(x)")

    title("sin(x) plotted for many values of x")

    plot(xdata,cosdata)
    ylabel("cos(x)")

    title("cos(x) plotted for many values of x")


    show()


if __name__ == "__main__":
    main()
  • 09.py
    from matplotlib.pyplot import figure, show, plot, grid, xlabel, ylabel, title, cla

from numpy import linspace, sin, cos

def main():
    #Building on the previous program, remove the first plot with cla()
    figure()

    xdata=linspace(0,10,100)
    sindata=sin(xdata)
    cosdata=cos(xdata)

    plot(xdata,sindata)
    grid()

    xlabel("x")
    ylabel("sin(x)")

    title("sin(x) plotted for many values of x")


    #cla() clears the current axis (only sin() is plotted there now)
    #cla() is most useful in a interactive session
    cla()

    plot(xdata,cosdata)
    xlabel("x")
    ylabel("cos(x)")

    title("cos(x) plotted for many values of x")


    show()


if __name__ == "__main__":
    main()
  • 10.py
    from matplotlib.pyplot import figure, show, plot, grid, xlabel, ylabel, title

from numpy import linspace, sin, cos

def main():
    #Building on the previous program, plot sin() and cos() in different places
    fig=figure()

    #Generate all the data that will be plotted
    xdata=linspace(0,10,100)
    sindata=sin(xdata)
    cosdata=cos(xdata)


    #Make a subplot for sin(x)
    fig.add_subplot(2,1,1)
    plot(xdata,sindata)
    grid()

    xlabel("x")
    ylabel("sin(x)")
    title("sin(x) plotted for many values of x")


    #Make a subplot for cos(x)
    fig.add_subplot(2,1,2)
    plot(xdata,cosdata)
    grid()

    xlabel("x")
    ylabel("cos(x)")
    title("cos(x) plotted for many values of x")


    show()


if __name__ == "__main__":
    main()
  • 11.py
    from matplotlib.pyplot import figure, show, plot, grid, xlabel, ylabel, title

from numpy import linspace, sin, cos

def main():
    #Building on the previous program, plot sin() and cos() in different places
    fig=figure()

    #Generate all the data that will be plotted
    xdata=linspace(0,10,100)
    sindata=sin(xdata)
    cosdata=cos(xdata)

    #Make a subplot for sin(x)
    fig.add_subplot(1,2,1)
    plot(xdata,sindata)
    grid()

    xlabel("x")
    ylabel("sin(x)")
    title("sin(x) plotted for many values of x")


    #Make a subplot for cos(x)
    fig.add_subplot(1,2,2)
    plot(xdata,cosdata)
    grid()

    xlabel("x")
    ylabel("cos(x)")
    title("cos(x) plotted for many values of x")


    show()


if __name__ == "__main__":
    main()
  • 12_Scatter.py
    from matplotlib.pyplot import figure, show, scatter

def main():

    fig=figure()

    #scatter plot two points - one at (1,10), the other at (2,5)
    #scatter() requires two arguments - a list of x coordinates and a list of y coordinates
    scatter([1,2], [10,5])

    show()



if __name__ == "__main__":
    main()
  • 12_Scatter_error.py
    from matplotlib.pyplot import figure, show, scatter

def main():

    fig=figure()

    #scatter plot two points - one at (1,10), the other at (2,5)
    #scatter() requires two arguments - a list of x coordinates and a list of y coordinates
    scatter([1,2], [10,5,10])

    show()



if __name__ == "__main__":
    main()
  • 13_Scatter.py
    from matplotlib.pyplot import figure, show, scatter
from numpy import random

def main():

    #Scatter a bunch of points taken from a Gaussian (normal) distribution

    coords=random.randn(1000,2)

    x=coords[:,0]
    y=coords[:,1]

    fig=figure()

    scatter(x,y)

    show()

if __name__ == "__main__":
    main()
  • 13_Scatter_alternate.py
    from matplotlib.pyplot import figure, show, scatter
from numpy import random, loadtxt

def main():

    #Scatter a bunch of points taken from a Gaussian (normal) distribution
    coords=loadtxt("normaldata.txt")

    x=coords[:,0]
    y=coords[:,1]

    fig=figure()

    scatter(x,y)

    show()


if __name__ == "__main__":
    main()
  • 14_Scatter.py
    from matplotlib.pyplot import figure, show, scatter, hist
from numpy import random

def main():
    #Scatter a bunch of points taken from a Gaussian (normal) distribution, and plot histograms for x and y data

    coords=random.randn(1000,2)

    x=coords[:,0]
    y=coords[:,1]


    fig=figure()

    #Each of the "add_subplot" calls make a different axis todraw things on.
    fig.add_subplot(3,1,1)
    scatter(x,y)

    fig.add_subplot(3,1,2)
    hist(x, bins=20)

    fig.add_subplot(3,1,3)
    hist(y, bins=20)

    show()


if __name__ == "__main__":
    main()
  • 15_Scatter.py
    from matplotlib.pyplot import axes, scatter, hist, axis, figure, show
from numpy import random

def main():
    #Build on the previous program and move the histograms to where they make sense

    coords=random.randn(1000,2)

    x=coords[:,0]
    y=coords[:,1]


    fig=figure()

    #How to create an axes: axes([left, bottom, width, height])
    #These three axes should not overlap - each one will plot something different.
    ax=axes([.05,.25,.7,.7])
    ax1=axes([.05,.05,.7,.15])
    ax2=axes([.8,.25,.15,.7])

    #Make a scatter plot, and then plot histograms for each dimension of data.
    ax.scatter(x,y)
    ax1.hist(x,bins=20)
    ax2.hist(y,bins=20, orientation="horizontal")

    show()


if __name__ == "__main__":
    main()
  • 16_Scatter.py
    from pylab import axes, scatter, hist, axis, figure, show
from numpy import random

def main():
    """
    Plot a normal and uniform set of synthetic data.
    """

    uniformplot()
    normalplot()


    show()


def uniformplot():
    #First, create a bunch of "experimental data"
    coords1=random.randint(0,1000,size=(1000,2))

    #Make a figure that we can plot on.
    fig=figure()

    #How to create an axes: axes([left, bottom, width, height])
    #These three axes should not overlap - each one will plot something different.
    ax=axes([.05,.25,.7,.7])
    ax1=axes([.05,.05,.7,.15])
    ax2=axes([.8,.25,.15,.7])

    x=coords1[:,0]
    y=coords1[:,1]

    #Make a scatter plot, and then plot histograms for each dimension of data.
    ax.scatter(x,y)
    ax1.hist(x,bins=20)
    ax2.hist(y,bins=20, orientation="horizontal")

    #Reset the scatter plot axis limits to "fit"
    ax.axis([0,1000,0,1000])


def normalplot():

    coords2=random.randn(1000,2)

    fig=figure()
    bx=axes([.05,.25,.7,.7])
    bx1=axes([.05,.05,.7,.15])
    bx2=axes([.8,.25,.15,.7])

    bx.scatter(coords2[:,0],coords2[:,1])
    bx1.hist(coords2[:,0],bins=20)
    bx2.hist(coords2[:,1],bins=20, orientation="horizontal")


if __name__ == "__main__":
    main()
  • 17_Pie.py
    from matplotlib.pyplot import figure, show, pie


def main():
    #Make a basic pie chart
    figure()

    piedata=[10,20,30]

    pie(piedata)

    show()


if __name__ == "__main__":
    main()
  • 18_Pie.py
    from matplotlib.pyplot import figure, show, pie


def main():
    #Build on the previous program and add labels to the pie pieces
    figure()

    piedata=[10,20,30]

    pie(piedata, labels=["First", "Second", "Third"])

    show()


if __name__ == "__main__":
    main()
  • 18_Pie_error.py
    from matplotlib.pyplot import figure, show, pie


def main():
    #Build on the previous program and add labels to the pie pieces
    figure()

    piedata=[10,20,30]

    pie(piedata, labels=["First", "Second"])

    show()


if __name__ == "__main__":
    main()

LATEST ANNOUNCEMENTS

Intro to Hadoop on Gordon
On January 31, the San Diego Supercomputer Center offers a no-fee introductory training workshop on Hadoop and its use in scientific and data intensive computing.

XSEDE Allocation Submissions Open
XSEDE is accepting Research Allocation Requests for the allocation period, April 1, 2013 to March 31, 2014. The submission period ends January 15, 2013.

OpenACC GPU Programming Workshop
The Pittsburgh Supercomputing Center, the National Institute for Computational Sciences and Georgia Tech will present an OpenACC GPU programming workshop on January 15 and 16, 2013.